Harold Abelson Deposition, in MPAA v. 2600

MA; July 5, 2000

See related files:
http://www.eff.org/IP/Video (EFF Archive)
http://cryptome.org/cryptout.htm#DVD-DeCSS (Cryptome Archive)
http://www.2600.com/dvd/docs (2600 Archive)
http://eon.law.harvard.edu/openlaw/dvd/ (Harvard DVD OpenLaw Project)


                                                      VOLUME: I
                                                      PAGES: 1-95
                                                      EXHIBITS: 1-5
                               UNITED STATES DISTRICT COURT
                               SOUTHERN DISTRICT OF NEW YORK
                               Case No. 00 CIV 277 (LAK)(RLE)   
                     UNIVERSAL CITY STUDIOS, INC.;    ) 
                     PARAMOUNT PICTURES CORPORATION;  ) 
                     METRO-GOLDWYN-MAYER STUDIOS,     ) 
                     INC.; TRISTAN PICTURES, INC.;    )
                     COLUMBIA PICTURES INDUSTRIES,    ) 
                     INC.; TIME WARNER ENTERTAINMENT  )
                     CO., L.P.; DISNEY ENTERPRISES,   ) 
                     INC.; AND TWENTIETH CENTURY FOX  )     
                     FILM CORPORATION,                ) 
                                Plaintiffs,           )  
                        VS.                           ) 
                     SHAWN C. REIMEREDES; ERIC CORLEY )
                     A/K/A "EMMANUEL GOLDSTEIN";      ) 
                     ROMAN KAZAN AND 2600 ENTERPRISES,) 
                     INC.                             ) 
                                Defendants.           )         
                                     DEPOSITION OF HAROLD ABELSON,
                     a witness called on behalf of the Plaintiffs, 
                     taken pursuant to the provisions of the Federal 
                     Rules of Civil Procedure, before Loretta 
                     Hennessey, Registered Merit Reporter and Notary 
                     Public in and for the Commonwealth of 
                     Massachusetts, at the offices of Choate, Hall & 
                     Stewart, 53 State Street, Boston, Massachusetts, 
                     on Monday, July 5, 2000, commencing at 10:49 a.m.


             1       APPEARANCES:
                        PROSKAUER ROSE LLP
             3               (By William M. Hart, Esq.)
                             1585 Broadway
             4               New York, New York  10036
                             for the Plaintiffs.
                        FRANKFURT GARBUS KLEIN & SELZ, P.C.
             7               (By Martin Garbus, Esq.)
                             488 Madison Avenue
             8               New York, New York  10022
                             for the Defendants.



             1                              I N D E X
             2       Witness        Direct  Cross  Redirect  Recross 
             3       HAROLD ABELSON
             4       (By Mr. Hart)   4                82
                     (By Mr. Garbus)         44               --
                                       E X H I B I T S
                     Exhibit No.                                Page  
                        1       Vita of Harold Abelson.          6
                        2       Notice of Deposition.           10
                        3       Subpoena.                       11
                        4       Declaration of Harold
            13                  Abelson.                        17
            14          5       Declaration of Michael
                                Shamos.                         61
            17                  _______________



             1                  P R O C E E D I N G S
             2                  HAROLD ABELSON, Sworn
             3       a witness called on behalf of the Plaintiffs, 
             4       having been duly sworn, was examined and 
             5       testified as follows:
             6                  DIRECT EXAMINATION
             7       BY MR. HART:
             8   Q.  How shall I address you?  Doctor?  Professor?
             9   A.  How about Hal.
            10   Q.  Thank you. 
            11                  Have you ever been deposed before?
            12   A.  No.
            13   Q.  Do you have an up-to-date curriculum vitae that 
            14       you could provide to me? 
            15   A.  Yeah. 
            16                  (Document exhibited to counsel.)
            17   Q.  Thank you. 
            18                  Are you currently employed?
            19   A.  Yes.
            20   Q.  What do you do?
            21   A.  I teach at MIT.
            22   Q.  What do you teach?
            23   A.  Computer science.
            24   Q.  Is that a full-time job?



             1   A.  Yes.
             2   Q.  Do you also consult or do anything else in 
             3       addition to teaching full time?
             4   A.  Yes, I consult.  I do most of my consulting right 
             5       now for Hewlett Packard.
             6   Q.  In what areas or area?
             7   A.  I work on Internet technology and some of their 
             8       electronic publishing.
             9   Q.  When you say "Internet technology," could you 
            10       briefly describe what that means? 
            11   A.  It has to do with how to make an infrastructure 
            12       that lets people mount services on the Internet; 
            13       also some more specific technology to make very 
            14       efficient Internet servers.
            15   Q.  And what particular areas of experience, 
            16       knowledge or expertise do you contribute to the 
            17       area that you just described? 
            18   A.  For MIT or for --
            19   Q.  No, I'm sorry, in your consulting capacities.
            20   A.  In my consulting?  I was part of organizing 
            21       Hewlett Packard's Internet Technology Division, 
            22       part of their general management of what kinds of 
            23       projects they'll pick, what kinds of, what kinds 
            24       of services they'd like to develop, and general 



             1       kinds of discussions about what kinds of research 
             2       Hewlett Packard Laboratories ought to be doing.
             3   Q.  Let me just get this straight.  I'm trying to 
             4       clarify this, so I apologize if I get it wrong. 
             5                  Is your consulting at Hewlett Packard 
             6       Hewlett Packard to enable Hewlett Packard to put 
             7       up material on a web site to promote what Hewlett 
             8       Packard does, or is it to enable Hewlett Packard 
             9       to offer technology and/or services to others in 
            10       connection with the delivery of content onto the 
            11       web?
            12   A.  Mostly the second. 
            13                  MR. HART:  Let's mark this, if you 
            14       don't mind, Exhibit 1. 
            15                  (Document marked as Exhibit 1
            16                  for identification.)
            17   Q.  Again, I apologize in advance if I misstate it, 
            18       but in connection with the consulting activities 
            19       you do for Hewlett Packard, does that involve the 
            20       delivery of content via the Internet?
            21   A.  It involves making the infrastructure that would 
            22       let other people do it. 
            23   Q.  Okay.  And by "infrastructure," what are we 
            24       referring to?



             1   A.  We're referring to, first of all, the hardware 
             2       infrastructure, the network infrastructure, 
             3       indexing, directory searching, tracking, content 
             4       protection in some cases.
             5   Q.  And when you say "content protection," what do 
             6       you mean?
             7   A.  Things like authentication, making sure that only 
             8       people who are authorized to get at various web 
             9       sites can do it.
            10   Q.  What about encryption as a form of protection?
            11   A.  Some encryption.
            12   Q.  Is part of the goal generally in the area of 
            13       content delivery on the Internet to increase the 
            14       speed at which large files can be delivered?
            15   A.  Repeat that.
            16                  MR. HART:  Let me have her read it 
            17       back.  She may get it more correctly than I do if 
            18       I have to restate it. 
            19                  (Question read.)
            20   A.  Oh, I'm sorry.  Yes. 
            21   Q.  And is it fair to say that that's done through a 
            22       combination of compression technology and broader 
            23       bandwidth on the Internet?
            24   A.  Yes, compression technology, broader bandwidth, 



             1       different ways of arranging switching, sometimes 
             2       storing files in multiple locations.  There's all 
             3       sorts of things. 
             4   Q.  From your vantage point, could you describe how 
             5       you see the pace of developments in those areas?
             6                  MR. GARBUS:  I'd object unless you 
             7       feel --
             8                  THE WITNESS:  Excuse me?
             9                  MR. GARBUS:  I'd object unless you 
            10       feel that's an area of your expertise; if it is, 
            11       then so state.
            12                  MR. HART:  You've got to keep your 
            13       voice up, Marty, because I don't know if the 
            14       court reporter can hear you, I know I can't.  It 
            15       may just be that I'm deaf.
            16   A.  I couldn't characterize how fast it's getting 
            17       faster, but it's improving, mostly through 
            18       different kinds of network technologies.
            19   Q.  By "network technologies," can you just generally 
            20       describe what you mean? 
            21   A.  Mostly increased use of optical networking and 
            22       advanced kind of switching, but I'm not an expert 
            23       enough to characterize that in any detail.
            24   Q.  Expert in network design?



             1   A.  Network design. 
             2   Q.  Do you ever use DeCSS?
             3   A.  No.
             4   Q.  Do you know what it is?
             5   A.  Yes, I know.
             6   Q.  Can you give us a plain English description of 
             7       what it is?
             8   A.  My understanding is that it's an algorithm that 
             9       encrypts the Content Scrambling System that was, 
            10       that's used on videos. 
            11   Q.  When you say "an algorithm," just for us 
            12       nonengineering types in the audience, you're 
            13       referring to a formula?
            14   A.  A formula, a recipe, what you build a computer 
            15       program to make.
            16   Q.  And are you aware whether a computer program 
            17       embodying DeCSS algorithm, formula, recipe has in 
            18       fact been constructed?
            19   A.  I'm told it has.  I've never looked at one 
            20       myself.
            21   Q.  Do you regard there being a difference between an 
            22       algorithm and an executable utility?
            23   A.  A difference? 
            24   Q.  Yes. 



             1   A.  A difference for what purpose?  Certainly things 
             2       are the same and things are different.
             3   Q.  Okay.  Could you explain the differences between 
             4       an algorithm and an executive utility embodying 
             5       that algorithm?
             6   A.  An algorithm is generally described at different 
             7       levels of abstraction than computer utility.  
             8       Some algorithms you can translate quite directly 
             9       and automatically into computer utilities.  Some 
            10       algorithms are best expressed as computer code in 
            11       the first place.  And sometimes an actual program 
            12       will have other details that are different from 
            13       the algorithm itself.
            14   Q.  Have you examined the DeCSS algorithm?
            15   A.  No.
            16   Q.  Have you ever examined any form of executable 
            17       utility called DeCSS or embodying DeCSS?
            18   A.  No.
            19                  MR. HART:  I'm going to mark a couple 
            20       of things here.  Ms. Reporter, if you would, 
            21       that's No. 2. 
            22                  (Document marked as Exhibit 2
            23                  for identification.)
            24                  MR. HART:  That will be No. 3. 



             1                  (Document marked as Exhibit 3
             2                  for identification.)
             3                  MR. HART:  We're on the record for 
             4       just one moment. 
             5   Q.  And I'm reluctant to call you Hal, although you 
             6       seem like a very nice man, and I'm not 
             7       necessarily a formal person, but I'm going to 
             8       show you what's just been marked by the reporter 
             9       as Exhibits 2 and 3.  And my question to you, 
            10       after you've looked at them -- you can take a 
            11       moment to look through each of the pages if you 
            12       like -- my question is:  Have you ever seen any 
            13       of those documents before?  And in answering, 
            14       please indicate which document by exhibit number. 
            15                  (Documents exhibited to witness.)
            16   Q.  It may help if I give you the ones that have been 
            17       marked by the reporter because they have the 
            18       exhibit numbers on them. 
            19                  (Documents exhibited to witness.)
            20   A.  Okay.
            21   Q.  Have you seen either Exhibits 2 or 3 before?
            22   A.  I've seen 3 before.  3 is the subpoena I was 
            23       served with.  2 looks like it's the same, but I 
            24       haven't seen it before.



             1   Q.  Fair enough.  You can actually hand those back. 
             2                  (Documents exhibited to counsel.)
             3   Q.  Did you collect any documents to bring with you 
             4       here today?
             5   A.  No.
             6   Q.  Were you aware that the subpoena called for you 
             7       to produce certain documents as listed on the 
             8       Schedule A to the subpoena?
             9   A.  Yes.  There was nothing relevant.
            10   Q.  I'm sorry?
            11   A.  I didn't see anything relevant in the subpoena.
            12   Q.  What do you mean by that? 
            13   A.  It asked if I had written things about DeCSS or 
            14       articles about things.  I don't have any.
            15   Q.  Is it your testimony that you had nothing in your 
            16       possession or control that responded to any of 
            17       the areas in the Schedule A of the subpoena, 
            18       Exhibit 3?
            19   A.  Yes, that's my testimony.
            20   Q.  Okay.  It's difficult when you use lawyer terms 
            21       like "relevant."  I have to find out what you 
            22       mean.
            23   A.  Sorry. 
            24   Q.  How did you first get involved in this case?



             1   A.  I believe either Marty or Ed Hernstadt phoned me 
             2       and asked if they could use a copy of the 
             3       deposition I made in another case. 
             4   Q.  Okay.  You were deposed in another case?
             5   A.  I'm sorry, declaration.  I'm sorry. 
             6   Q.  And what other case are you referring to?
             7   A.  This is Bernstein versus U.S.
             8   Q.  And what was the gist of your testimony in 
             9       Bernstein versus U.S.?
            10   A.  It's the same as the declaration I've submitted 
            11       in this case:  It said that computer code is 
            12       expressive, and people use computer code as a 
            13       means of expression. 
            14                  Actually, I have it. 
            15   Q.  I'm sorry, just because the record is blind, it 
            16       didn't see you reaching into your bag and pull 
            17       out a document.
            18   A.  I'm pulling out a copy of the declaration I made.
            19   Q.  You now have in front of you a copy of the 
            20       declaration that you filed in this case?
            21   A.  In this case. 
            22   Q.  I prefer you put that to the side right now, if 
            23       you don't mind.
            24   A.  Okay.  (Witness complies.)



             1   Q.  We're going to go through that in a few minutes.  
             2       Thank you. 
             3                  Did you have any awareness of this 
             4       case prior to the time that you were contacted by 
             5       Mr. Garbus or Mr. Hernstadt?
             6   A.  Only a minor awareness that it was going on, but 
             7       I didn't know any of the details of it.
             8   Q.  What did you know before you heard from Mr. 
             9       Garbus and Mr. Hernstadt?
            10   A.  Let's see if I can remember. 
            11                  Well, I knew that some sort of 
            12       complaint had been filed having to do with DeCSS, 
            13       but I didn't know the details of it. 
            14   Q.  And were you informed in that first discussion, 
            15       whether it was a discussion with Mr. Garbus or 
            16       Mr. Hernstadt, were you informed more about the 
            17       case at that time during that discussion?
            18   A.  Only a little bit.  What I was told is my 
            19       declaration in the Bernstein case was relevant to 
            20       this, could they use it. 
            21   Q.  Did Mr. Garbus or Mr. Hernstadt describe to you 
            22       their view of the case?
            23   A.  No, not at that time.
            24   Q.  Did there come a time when they did?



             1   A.  I asked Marty about it this morning as we were 
             2       waiting.  We haven't discussed things much other 
             3       than that. 
             4   Q.  Apart from the first call that you got either 
             5       from Mr. Garbus or Mr. Hernstadt to ask you to 
             6       put in a declaration in this case, and your 
             7       meeting here today with Mr. Garbus, how many 
             8       other contacts or communications have you had 
             9       with anyone at the defendants' law firm?  And by 
            10       that I mean Mr. Garbus or Mr. Hernstadt or any of 
            11       their associates or colleagues.
            12   A.  Perhaps four.
            13   Q.  I'm sorry?
            14   A.  Perhaps four.  Only with Mr. Garbus and Mr. 
            15       Hernstadt, mostly having to do with scheduling 
            16       this deposition.
            17   Q.  Okay.  Were there any discussions concerning the 
            18       substance of the case?
            19   A.  We talked a little bit about computer code as 
            20       means of expression.  We talked a little bit this 
            21       morning about general issues of copyright, but 
            22       nothing in depth.
            23   Q.  Have you followed the legislative progress and 
            24       enactment of the Digital Millennium Copyright 



             1       Act?
             2   A.  Yes, I have. 
             3   Q.  Have you ever submitted any papers, views, 
             4       testimony or the like in connection with the 
             5       legislation, its enactment or any rule making 
             6       related to that law? 
             7   A.  Nothing formal.  On the other hand, I am a 
             8       constituent of Barney Frank's and I had 
             9       correspondence with him when he was on the 
            10       Judiciary Committee when that law was before the 
            11       Judiciary Committee.
            12   Q.  And this correspondence included your views about 
            13       the proposed enactment?
            14   A.  Yes, it did. 
            15   Q.  You say your full-time job is teaching about 
            16       computers and computer programs at MIT?
            17   A.  Teaching and research. 
            18   Q.  Can you generally describe for us what specific 
            19       areas within computers and computer programming 
            20       you teach or research in?
            21   A.  I do research in artificial intelligence and in 
            22       general to the interaction between the artificial 
            23       intelligence and computing for scientific and 
            24       engineering uses. 



             1                  I teach -- for many years I taught 
             2       MIT's major introductory programming course.  And 
             3       more recently I've been teaching a joint course 
             4       between MIT and Harvard Law School having to do 
             5       with the interaction between policy and 
             6       technology issues involving the Internet.
             7   Q.  Has this case come up in the course of that 
             8       class?
             9   A.  No. 
            10                  MR. HART:  Sorry? 
            11                  MR. GARBUS:  I didn't say anything. 
            12                  MR. HART:  I'd like to mark your 
            13       declaration now.  We're up to 4. 
            14                  (Document marked as Exhibit 4
            15                  for identification.)
            16   Q.  I show you what the reporter has just marked as 
            17       Exhibit 4 and ask you to identify that for us, 
            18       please. 
            19                  (Document exhibited to witness.)
            20   A.  Yes, that's my declaration.
            21   Q.  Thank you. 
            22                  MR. GARBUS:  You may want to keep it 
            23       in front of you.  He'll be asking you questions 
            24       about it. 



             1   Q.  Were there any prior drafts of this declaration; 
             2       that is, prior to the one that we see in front of 
             3       us marked as Exhibit 4?
             4   A.  There's the declaration I gave in the Bernstein 
             5       case.  I don't think there are any substantial 
             6       differences between them. 
             7   Q.  Between Exhibit 4 --
             8   A.  Between this one and the previous one.
             9   Q.  Between this one and the one you filed in 
            10       Bernstein?
            11   A.  Right.
            12   Q.  I notice that the penultimate page stops at 
            13       Paragraph 12, kind of in the middle of the page, 
            14       and there's a blank, and you turn to the next 
            15       page for signature.  Do you see what I'm saying?
            16   A.  No, there's nothing.... 
            17   Q.  In other words --
            18   A.  There was nothing extra there.  I think the only 
            19       difference is that when -- I'm not sure. I 
            20       believe when I did the Bernstein declaration I 
            21       was still teaching the course, and now I'm not 
            22       anymore, so Paragraph 3, I think, is, Paragraph 
            23       3, I think has changed, but I'm not positive on 
            24       that. 



             1   Q.  I'm sorry, but you're not?
             2   A.  But I'm not positive, I'm not even positive of 
             3       that. 
             4   Q.  Why aren't you teaching that course anymore?
             5   A.  I just -- people move on, teach different 
             6       courses, and I've done it since 1979.  It seemed 
             7       like.... 
             8   Q.  Are you getting paid anything for your 
             9       involvement in this case?
            10   A.  No.  I got a witness fee attached to this.
            11   Q.  That was from me.
            12   A.  That was from you?  Thank you.  I'm told the 
            13       going rate in Boston is $7.
            14   Q.  You're welcome. 
            15                  I understand you're going to be away 
            16       during the trial of this case; is that right?
            17   A.  I'm planning to be on the West Coast next week 
            18       and the week after.
            19   Q.  And how long have those plans been in place?
            20   A.  The plan for next week has been in place about a 
            21       month, and the plan for the week after is up in 
            22       the air at the moment.  It has to do with my 
            23       consulting with Hewlett Packard --
            24   Q.  I see. 



             1   A.  -- what I have to do there. 
             2   Q.  I see.  But until it became -- when did it become 
             3       up in the air?
             4   A.  I spend the summer going back and forth between 
             5       Cambridge and Palo Alto, and in general I'm not 
             6       sure which coast I'm going to be on which week.
             7   Q.  I see.  And were you ever informed by Mr. Garbus 
             8       or Mr. Hernstadt about the trial date in this 
             9       case?
            10   A.  I was, but I don't remember.  I thought you said 
            11       next week. 
            12   Q.  Do you recall when you were first informed by Mr. 
            13       Garbus or Mr. Hernstadt about the trial date in 
            14       this case?
            15   A.  One of our discussions about the depositions.  
            16       Probably sometime in the last three weeks, but 
            17       I'm not positive. 
            18   Q.  Okay.  What kind of Internet connection do you 
            19       have in your office?
            20   A.  In my office? 
            21   Q.  Uh-huh.
            22   A.  I have a hundred megabit per second Ethernet that 
            23       attaches through the MIT artificial intelligence 
            24       laboratory to MIT's major, what they call the 



             1       backbone network at MIT, which itself has roughly  
             2       a hundred megabit connection to what's called 
             3       NERnet.
             4   Q.  NERnet?
             5   A.  The New England Regional Network, I believe, and 
             6       from there to one of the major backbone networks.
             7   Q.  Do you have any knowledge about the type, speed 
             8       or bandwidth of the Internet connections that are 
             9       available to students at MIT generally, whether 
            10       in student facilities or dorm rooms or the like?
            11   A.  What I have is typical of MIT.  I think all of 
            12       our dormitory rooms have Ethernet connections.  I 
            13       don't remember whether all of them are hundred 
            14       megabit.  Some of them may be 10 megabit. 
            15   Q.  I'm going to ask you a couple of questions that 
            16       may require you to do some rough computations.
            17   A.  Oh, boy.  A test.
            18   Q.  I apologize in advance.
            19   A.  A test.
            20   Q.  No, it's not a test. 
            21                  With that type of Internet connection, 
            22       what's the effective transfer rate of, say, a one 
            23       gig file within that network?
            24   A.  Okay.  I'm going to embarrass myself.



             1   Q.  Please.  We've all done it. 
             2   A.  So a one gig file?  Gigabyte, I assume? 
             3   Q.  Yes. 
             4   A.  So that's 8 gigabits.  So 8 times 10 to the 10th 
             5       megabits. 
             6                  You want to know how long to transfer 
             7       a one gigabyte file?
             8                  Ten megabits per second. 
             9                  MR. GARBUS:  Are you factoring the 
            10       time of day of this transmission?  Midnight?
            11                  MR. HART:  Please, Marty.  Whether you 
            12       were saying that in jest or not, I would 
            13       appreciate you confining your comments at this 
            14       point.
            15                  MR. GARBUS:  It was in jest. 
            16   A.  That's 8,000 megabits, roughly.  One megabit per 
            17       second, that would be 8,000 seconds; ten, that 
            18       would be 800 seconds; a hundred megabits would be 
            19       80 seconds.  So if I take ten megabits per 
            20       second, that's -- a hundred megabits per second, 
            21       which is roughly a minute and a half for a one 
            22       gigabyte file.  I hope I'm getting this right.
            23   Q.  Okay. 
            24   A.  A ten gigabyte file? 



             1   Q.  No, one gigabyte file.
             2   A.  Looks to me to be at a hundred megabits a second, 
             3       it's a minute and a half; and at ten megabits per 
             4       second, it's about 15 minutes. 
             5                  That's theoretical capacity.
             6   Q.  Right.
             7   A.  Now, state the question again so I can understand 
             8       the assumptions.  That's theoretical capacity.
             9   Q.  I understand.
            10   A.  If I were doing it at MIT, for example.
            11   Q.  Right.
            12   A.  So if I were doing it at MIT, MIT's backbone as a 
            13       whole would tolerate a hundred megabits per 
            14       second, so if there were ten people trying to use 
            15       the network, I'd get maybe a 10th of that.
            16   Q.  Is that always true of network connections, or 
            17       does that depend on how they're switched?
            18   A.  It depends on how they're switched, and it 
            19       depends on just lots and lots and lots of 
            20       details.
            21   Q.  So let's say on your average day, taking into 
            22       account the humidity and the other factors Marty 
            23       was alluding to, in the real world, what transfer 
            24       rate are you getting on a hundred meg network at 



             1       MIT?  What's your real yield, roughly?
             2   A.  I would have to guess around 25. 
             3                  You have to understand, for very short 
             4       periods of time, I can get close to the maximum.  
             5       If you're trying to do something over a long 
             6       period of time, you're sharing the connection.
             7   Q.  What do you call a long period of time in the 
             8       Internet world?
             9   A.  A minute. 
            10   Q.  Taking our hypothetical one gig file and 
            11       factoring real world factors into it, what do you 
            12       think is a real world case for the transfer of 
            13       that one gig file over your hundred meg network 
            14       at MIT?
            15   A.  I'd have to guess around 15 or 20 minutes. 
            16   Q.  Are you familiar with how large a feature length 
            17       movie file is as embodied on a DVD?
            18   A.  Not really.
            19   Q.  Do you have any anecdotal information?
            20   A.  Let me think.  DVD is 600 megabytes.  DVD -- no, 
            21       I just don't.  Something like -- it's multiple 
            22       gigabytes, but....
            23   Q.  I'm not asking you to guess, I just wanted to 
            24       know what you knew. 



             1                  Are you familiar with video 
             2       compression technologies, generally?
             3   A.  Only a little bit. 
             4   Q.  Do you have any idea --
             5                  MR. GARBUS:  Bill, do you want to 
             6       stipulate as to a general number in case you want 
             7       the doctor to figure it out?  Just take a number 
             8       that we can agree is kind of an average or 
             9       working number so he can do the rest of the 
            10       computation?
            11                  MR. HART:  No. 
            12   Q.  We're not going to do a whole lot more math here, 
            13       Doctor.  But I thank you for that computation a 
            14       minute ago. 
            15                  Do you have any knowledge of the kinds 
            16       of compression ratios that are currently 
            17       available under, for example, the MPEG-4 standard 
            18       for video?
            19   A.  I would guess, this is just a guess, a factor of 
            20       ten, but it depends on quality.
            21   Q.  Is that also something that is developing as a 
            22       science or technology, that is, there are 
            23       improvements being wrought on a continual basis? 
            24   A.  Well, I mean, MPEG-4 is fairly recent.  I'm not 



             1       aware of things that are going on beyond that, 
             2       but I know there are people working on it. 
             3   Q.  This is just part of the nature of this beast, 
             4       because the record gets cluttered.  I'll try not 
             5       to interrupt you. 
             6                  How do you know about MPEG-4?
             7   A.  Sort of general reading in the technology 
             8       literature. 
             9   Q.  To your knowledge, is it a commonly available 
            10       piece of software?
            11   A.  I don't know.  I know it's used. 
            12   Q.  Have you ever heard of Napster?
            13   A.  Oh, yeah.
            14   Q.  How did you hear of it?
            15   A.  More general reading both in technology 
            16       literature and newspapers; read an article about 
            17       it this morning in the Wall Street Journal while 
            18       I was waiting. 
            19   Q.  Do you know if MIT has had a problem with Napster 
            20       in terms of student usage?
            21   A.  No.  MIT's noticed it, and I know that other 
            22       universities have.  MIT in general has enough 
            23       capacity that Napster use wasn't a problem, 
            24       although it was at the level where if it were 



             1       much more, it would become a serious problem.
             2   Q.  And that's because the bandwidth was being used 
             3       for so-called file sharing of MP3 audio files?
             4   A.  Yes.  Just because MIT has a lot of internal 
             5       capacity on its network.
             6   Q.  What kind of Internet hook-up do you have at 
             7       home?
             8   A.  Connection through Media One.
             9   Q.  What's the effective bandwidth through that?
            10   A.  I believe it's 300 kilobits upstream, and a 
            11       little over a megabyte downstream.
            12   Q.  And by "upstream" and "downstream," you mean that 
            13       you could get a megabit in if you were 
            14       downloading something from the Net?
            15   A.  You could get a megabit -- well, again, 
            16       theoretically.
            17   Q.  Right.  I just wanted to define upstream and 
            18       downstream.
            19   A.  Yes, you can get stuff from the Internet about 
            20       three or times as fast as you could upload it. 
            21   Q.  Gotcha. 
            22                  CONFIDENTIAL
            24   Q.  Do you have any view of what the sort of standard 



             1       bandwidth is that's available at major 
             2       universities and colleges throughout the country?
             3   A.  I tend to think it's -- are you talking about 
             4       bandwidth in the university or inside the 
             5       university to the outside?
             6   Q.  Let's do them both. 
             7   A.  From is usually either 10 or a hundred meg.
             8   Q.  Meg?
             9   A.  Megabits per second.  Internally, it's about the 
            10       same, 10 or a hundred.  Again, with the external 
            11       connections, there tends to be bottlenecks and 
            12       things. 
            13   Q.  Just so that I'm clear, you may have had it 
            14       perfectly in your own mind, were we speaking of 
            15       the 10 or 100 meg that we were speaking of when 
            16       we were talking about MIT a few minutes ago, 
            17       megabytes?
            18   A.  Megabits per second.
            19   Q.  As opposed to megabytes?
            20   A.  Right.  We're talking networks, you usually talk 
            21       about megabits. 
            22   Q.  So just to be clear, again -- it may be my fault 
            23       in terms of how I ask the question, I apologize 
            24       for that -- when we were talking about what the 



             1       typical network bandwidth is at major colleges 
             2       and universities throughout the country 
             3       internally --
             4   A.  Right.
             5   Q.  -- what is that bandwidth rate, approximately in 
             6       your view?
             7   A.  No, my guess, it's either 10 or a hundred.
             8   Q.  10 or a hundred what?
             9   A.  Oh, I'm sorry.  I'm sorry.  10 or a hundred 
            10       megabits per second.
            11   Q.  Thank you. 
            12   A.  Theoretical maximum transfer. 
            13   Q.  Okay.  And that theoretical is subject to the 
            14       same variables you described a few minutes ago 
            15       when we were talking about MIT's network, right?
            16   A.  Right.  The way to understand that is that 
            17       maximum transfer rate means that's as much as you 
            18       have for everyone who might be trying to do this 
            19       at one particular time at the university. 
            20   Q.  But doesn't that also depend on how it's 
            21       switched, that is, how the network is switched?
            22   A.  Yes, it depends in complicated ways that I'm not 
            23       expert enough to really characterize.
            24   Q.  I'm sorry, I didn't mean to interrupt you.  In 



             1       general terms, do you have any understanding of 
             2       the difference between a shared network and a 
             3       switched network?
             4   A.  A shared network and a switched network?  No, I 
             5       don't.
             6   Q.  Are there networks, to your knowledge, that can 
             7       make the bandwidth that you mentioned, let's say, 
             8       10 meg, available to all users no matter how many 
             9       users are on it? 
            10   A.  Not that I'm aware of. 
            11   Q.  Okay.  Can you read object code?
            12   A.  Sure.
            13   Q.  Is there a difference, in your view, in the 
            14       expressive content between object code and source 
            15       code?
            16   A.  It depends very much on the circumstances.  In 
            17       general, each one will express different things.
            18   Q.  But you have not looked at either the object code 
            19       or source code versions of DeCSS?
            20   A.  No. 
            21                  MR. GARBUS:  Can you take a second?  I 
            22       just want to make one call. 
            23                  MR. HART:  Let the record reflect Mr. 
            24       Garbus is making a phone call.  I don't think 



             1       there's a question pending right now.  We can 
             2       take five. 
             3                  (Discussion off the record.)
             4                  MR. HART:  Okay.  We're back. 
             5                  (Question and answer read.)
             6   Q.  What expressive value does object code have?
             7   A.  Well, object code can tell me a tremendous amount 
             8       about efficiency.  Source code is generally 
             9       written to be machine independent.
            10   Q.  Source code is?
            11   A.  Source code is.  And then when you put that 
            12       through a translator to translate it into 
            13       something for a specific machine, then the object 
            14       code is telling you about, in general, the 
            15       interactions of the original source program with 
            16       specific details of the architecture of the 
            17       machine.
            18   Q.  The object code is?
            19   A.  That's what the object code will tell you.  Then 
            20       for some applications that will be absolutely 
            21       critical. 
            22   Q.  In other words, to see how efficiently the 
            23       program runs, object code has expressive value?
            24                  MR. GARBUS:  I object to what you 



             1       said.  He said more than that.  But go ahead. 
             2   A.  It's a little bit more than efficiency.  Again, 
             3       it's interactions of the detailed, details of 
             4       this program execution with the specific 
             5       architecture of the machine, and that can be not 
             6       only efficiency but things like reliability, 
             7       things like how it will interact with maybe other 
             8       programs that are running on the machine at the 
             9       same time. 
            10   Q.  What, to your knowledge, was the issue in dispute 
            11       in the Bernstein case?
            12   A.  The issue in the Bernstein case was that the 
            13       government had claimed that posting of various 
            14       kinds of encryption algorithms, or even in the 
            15       Bernstein case, a technical paper describing 
            16       them, violated the Export -- initially it was 
            17       ITAR, but later the Export Control Act. 
            18   Q.  What is your understanding of the Export Control 
            19       Act?  I'm not looking for a legal dissertation 
            20       here.
            21   A.  My understanding is that it's horribly 
            22       complicated. 
            23   Q.  Is it fair to say until recently there were 
            24       restrictions on the bit length of certain 



             1       encryption codes that were put into place by 
             2       government regulation?
             3   A.  There were certain restrictions on bit length.  
             4       There were other restrictions as well, but I 
             5       can't remember them.  It's a very convoluted and 
             6       complicated act, as I recall. 
             7   Q.  Do you remember the bit length restriction that 
             8       was in place at the time of the Bernstein case?  
             9       Was it a 40-bit limitation?
            10   A.  At one point I believe it was 40 bits.  I think 
            11       at some point it got changed to 56.
            12   Q.  Okay.  Have you ever gone to the 2600 web site?
            13   A.  Yes, I did, when I was looking at this case.  
            14       Eventually I read it because 2600 had a 
            15       discussion of the case, so I looked at it, but 
            16       just only briefly.
            17   Q.  Was that before or after you got the call from 
            18       Messrs. Garbus and Hernstadt?
            19   A.  After.
            20   Q.  I'm just having to complete the record.
            21   A.  Yes, it was after. 
            22   Q.  Did you ever have occasion to read the judge's 
            23       opinion or opinions in this case involving the 
            24       preliminary injunction?



             1   A.  I skimmed them, again, recently.  I skimmed them. 
             2   Q.  Now, if I can try and sum up your declaration in 
             3       one sentence.
             4                  MR. GARBUS:  That might be a problem. 
             5                  MR. HART:  It might be, but let me try 
             6       it. 
             7   Q.  Would it be fair to say that it's your view that 
             8       computer programs are expressive in nature and 
             9       convey ideas that are useful for people that are 
            10       involved in the evaluation or study of computer 
            11       programs?
            12                  MR. GARBUS:  I'll object.  I think he 
            13       didn't say it that narrowly.
            14                  MR. HART:  Okay. 
            15                  MR. GARBUS:  Why don't you let the 
            16       witness --
            17   Q.  You can correct it.  I'm trying to move it along.
            18                  MR. GARBUS:  Why doesn't the witness 
            19       say what he intends to say in his affidavit.
            20                  MR. HART:  You're going to object that 
            21       the declaration will speak for itself.
            22                  MR. GARBUS:  Why don't you just tell 
            23       us.... 
            24   A.  I guess it's my declaration that computer 



             1       programs are mediums of expression for many 
             2       purposes, and they convey ideas.  Often a 
             3       computer program will be the most appropriate way 
             4       to express a particular idea about how things are 
             5       done, and it's quite independent of any 
             6       particular purpose that someone wants to use it 
             7       for. 
             8                  MR. GARBUS:  Could you just mark that 
             9       answer because I'm going to want to go back to 
            10       it. 
            11   Q.  When you said at the end of your answer, in rough 
            12       ideas, independent of the use to which the 
            13       program can be put --
            14   A.  Yes.
            15   Q.  -- could we fit into that general statement, 
            16       independent, say, for example, of the use of a 
            17       decryption utility? 
            18   A.  Wait. 
            19                  MR. GARBUS:  I object to the question.  
            20       Too many independents.
            21   A.  Try again.  I got lost on you.
            22   Q.  Okay.  I'm sorry.  That was my fault, I 
            23       apologize. 
            24                  Even though you haven't examined or 



             1       analyzed DeCSS in either source or object code 
             2       form --
             3   A.  Right, that's correct.
             4   Q.  -- is it your position that as a general 
             5       proposition DeCSS must have some expressive value 
             6       in its object or source code iterations?
             7   A.  I'm not sure about "must have," but I would 
             8       certainly expect that it does.
             9                  MR. GARBUS:  Do you have a copy of the 
            10       codes with you?  Maybe we can show it to the 
            11       witness.
            12                  MR. HART:  You might want to do that.  
            13       I'll continue my line of questioning, Marty, 
            14       though, if you don't mind. 
            15   Q.  I believe in your previous answer, which I made a 
            16       garbled question over, and I apologize for that, 
            17       I believe you said that a program can convey 
            18       ideas independent of whatever use it's for?
            19   A.  Yes.
            20   Q.  I think we should go back -- is that right?
            21   A.  Yes, that's right. 
            22   Q.  So I guess this is a syllogism.  You deal with 
            23       this more than I do.
            24   A.  Okay.



             1   Q.  Does that mean that DeCSS, as a decryption 
             2       utility, would, in your view, have value to 
             3       communicate ideas totally independent of its use 
             4       as a decryption utility? 
             5   A.  Oh, yes, certainly.
             6   Q.  Is that essentially your testimony?
             7   A.  I can give you an example.  It's CSS was 
             8       incompetently designed.  So that's an idea that 
             9       you can get from looking at DeCSS, I would 
            10       assume.
            11   Q.  It's also an idea you could express just in the 
            12       manner you did, isn't it? 
            13   A.  (Witness nods.)
            14                  MR. GARBUS:  Objection.
            15   Q.  You better verbalize --
            16   A.  It would say more.  Yes.  It would tell me more 
            17       details about that particular kind of 
            18       incompetence.  And, sure, there are other ways 
            19       one could say that, but often computer code is 
            20       the most succinct way to do that. 
            21   Q.  Now, let me pose this to you, because I want to 
            22       see where you come out in terms of potential for 
            23       abuse.
            24   A.  Abuse? 



             1   Q.  Yeah.  Let's take --
             2                  MR. GARBUS:  He's not abusive.
             3                  THE WITNESS:  I'm not sure who's being 
             4       abused here.
             5                  MR. HART:  Well, I hope no one. 
             6   Q.  Let's take some code that in the form of an 
             7       executable utility would enable anyone who 
             8       downloads it from the Internet to walk into 
             9       CitiBank and cause cash to be spit out of 
            10       someone's account.
            11   A.  Okay. 
            12   Q.  And I guess I need your view as to where you 
            13       balance the interests in terms of interest and 
            14       information about the program versus the 
            15       proliferation of same on the Internet when that 
            16       risk exists.
            17                  MR. GARBUS:  I would object.  I would 
            18       allow the witness to answer, if you -- it seems 
            19       to me that's an endless dialogue, but go ahead.
            20   A.  Just ask again so I can --
            21   Q.  Sure.  Here's the example:  Someone develops a 
            22       utility that if downloaded from the Internet 
            23       enables you to basically go to any CitiBank ATM 
            24       machine and take money out of someone else's 



             1       account. 
             2   A.  Right. 
             3   Q.  And in view of what I've said about the sort of 
             4       expressive content of computer code and interest 
             5       in examining that for whatever reason --
             6   A.  Right. 
             7   Q.  -- I guess I'm asking you, Hal Abelson, where you 
             8       come out professionally and personally on the 
             9       issue of promulgating that executable utility on 
            10       the Internet for academic instructive value if 
            11       there is the risk of abuse that people could 
            12       download it and take money out of CitiBank using 
            13       it.  Do you understand my question? 
            14   A.  I'd say in that particular case, I would tell 
            15       CitiBank about it, wait a couple of days, and 
            16       then make it very clear that either I or someone 
            17       else would promulgate that code.  I think that's 
            18       probably the most -- I would argue that's 
            19       probably the most useful thing to do for 
            20       CitiBank.  But, again, it's kind of up to them to 
            21       say.
            22   Q.  Based on the assumption that they could change 
            23       their system?
            24   A.  Based on the assumption either that they could 



             1       change their system or they certainly, they 
             2       certainly should know about that weakness, and 
             3       they certainly should know that that would be a 
             4       really serious weakness.
             5   Q.  And notifying CitiBank about that weakness is 
             6       different than promulgating that utility on the 
             7       Internet, yes?
             8   A.  It's different.  Sometimes it's very difficult to 
             9       get people to change, it's very difficult for the 
            10       people who you inform of this thing to make 
            11       changes happen.
            12   Q.  Do you have a view as to the expressive value of 
            13       promulgating DeCSS on the Internet in this case?
            14   A.  View as to the expressive value? 
            15                  MR. GARBUS:  I object to the form of 
            16       the question, but I'll certainly let Professor 
            17       Abelson answer.  He's already testified --
            18                  MR. HART:  Please. 
            19                  MR. GARBUS:  Go ahead. 
            20   A.  Well, I think it does quite well at expressing 
            21       the nature of the Content Scrambling System.  I 
            22       think it -- I would venture to say that had the 
            23       people who designed the Content Scrambling System 
            24       had the benefit of the kinds of things I've read 



             1       on the Internet about DeCSS, they certainly would 
             2       have designed a different algorithm.
             3   Q.  You've read things about DeCSS on the Internet; 
             4       is that right?
             5   A.  Yes.
             6   Q.  And despite the fact that you're testifying in 
             7       this case as an expert, and despite your 
             8       expertise in computers and software, you've had 
             9       no occasion to actually look at the source or 
            10       object code in connection with your evaluations 
            11       or your opinion or your study?
            12   A.  I wasn't asked to do that.
            13   Q.  I see.  But you didn't find it necessary as a 
            14       professional to do that? 
            15   A.  Necessary for what? 
            16   Q.  To understand DeCSS or its expressive value, all 
            17       the things you testified to.
            18   A.  Oh, in general?
            19   Q.  Yes.
            20   A.  I'm not particularly looking at those kind of 
            21       encryption systems.
            22   Q.  But we're discussing that kind of encryption 
            23       system in this case, are we not? 
            24   A.  Well, you are.  I'm testifying about the general 



             1       expressiveness of code.
             2   Q.  Okay.  And I'm simply confirming with you, sir, 
             3       that in the course of your testifying about the 
             4       expressive nature of code, and in the course of 
             5       your reviewing things on the Internet about 
             6       DeCSS, you, as a professional, as an expert in 
             7       this area, saw no need to actually look at the 
             8       code, did you?
             9   A.  It's one of the very many algorithms I haven't 
            10       looked at. 
            11   Q.  Is there anything else that you intend to testify 
            12       to in this case in connection with DeCSS, CSS, or 
            13       any of the issues as you understand them in the 
            14       case?
            15   A.  That I intend to testify about?  I didn't come 
            16       here with intentions to testify.  I'm just 
            17       responding to both of your questions.
            18   Q.  I see.  Were there areas that were identified by 
            19       Mr. Garbus or Mr. Hernstadt, or anyone else, for 
            20       that matter, as to topics for you to testify on 
            21       in this case?
            22   A.  No.
            23   Q.  Do you have a web site?
            24   A.  Do I have a web site?  My research group at MIT 



             1       has a web site.
             2   Q.  What's that group's web site called?  If I were 
             3       to go look for it, how would I find it?
             4   A.  The project on mathematics and amorphous 
             5       computer -- I'm sorry.  The MIT project on 
             6       mathematics and computing.
             7                  MR. HART:  You know what might make 
             8       sense to expedite this, if you give me about ten 
             9       minutes, I'm see what I have left, it may enable 
            10       me to consolidate.
            11                  MR. GARBUS:  I'm going to ask some 
            12       questions, so what you may want to do, it might 
            13       make things go faster, after I ask questions, I 
            14       presume you're going to want to ask some also.  
            15       If you want, I can go now, or we can do as you 
            16       wish; namely, give you a few minutes for you to 
            17       do it.
            18                  MR. HART:  Why don't we give me a few 
            19       minutes, then we'll --
            20                  MR. GARBUS:  Do you have with you 
            21       copies of the DeCSS source and object code?  If 
            22       so --
            23                  MR. HART:  I don't think so. 
            24                  MR. GARBUS:  We'll take ten minutes. 



             1                  (Brief recess.)
             2       BY MR. HART: 
             3   Q.  Have you prepared any kind of report or summary 
             4       of your views or possible testimony in connection 
             5       with this case? 
             6   A.  No. 
             7                  MR. HART:  I don't think I have 
             8       anything more for you, Dr. Abelson.  I thank you 
             9       for your time and for your candor
            10                  CROSS EXAMINATION
            11       BY MR. GARBUS:
            12   Q.  Dr. Abelson, on your site do you link to the site 
            13       of David Touretzky?
            14   A.  Yes, I put a link to that site after I noticed it 
            15       when I was looking at the various things about 
            16       the DeCSS.
            17   Q.  Do you know whether the Touretzky site has the 
            18       object or source code for DeCSS?
            19                  MR. HART:  Object, form. 
            20   A.  I don't.  I don't know.  I know it had a whole 
            21       bunch of different forms of DeCSS.
            22   Q.  Now, before you used the express "architecture of 
            23       the machine."  Can you tell me what you mean by 
            24       that?  You were talking about how -- do you 



             1       recall your reference to that?
             2   A.  Yes. 
             3                  Well, in general it's the 
             4       architecture, it's the structure by which the 
             5       machine is put together.  So it's such things as 
             6       the number of registers that there will be, the 
             7       number of bits in the data paths that connect 
             8       them, the number and kinds of operations that can 
             9       happen simultaneously, the size of various 
            10       registers and portions of memory.
            11   Q.  And that, you indicated before, was communicated 
            12       by object code; is that right?
            13   A.  Yes.  Well, you'll see that immediately 
            14       reflected.  I should say that differently.  A 
            15       piece of source code could potentially translate 
            16       into many different kinds of object code.  And 
            17       then depending on the sophistication of the 
            18       computer, you'll try to take advantage of 
            19       different elements of the machine, and you'll see 
            20       that reflected in the object code.
            21   Q.  When you say source code can be translated to 
            22       different object codes, can you describe that for 
            23       me?
            24   A.  Sure.  There's a translation process.  In general 



             1       there are many levels in which languages will be 
             2       translated.  So in the very simplest case, you'll 
             3       have a program that takes in source code as 
             4       prepared by a human programmer and produce a code 
             5       called ASSEMBLY code, which is rather close to 
             6       the machine, although it's symbolic.  And then 
             7       ASSEMBLY code will be further translated by 
             8       something called an assembler to something that's 
             9       really at the level of the machine. 
            10                  In more general cases, you'll have a 
            11       whole bunch of levels of intermediate code; and 
            12       even more generally, a program can be prepared in 
            13       a multiplicity of languages.  So it would be -- 
            14       it's often possible to write much of an algorithm 
            15       in a very high level language, then write 
            16       critical details, in fact dropping down to the 
            17       ASSEMBLY level.
            18   Q.  When you say a multiplicity of languages, can you 
            19       explain that for me? 
            20   A.  Yeah, a good example would be something like 
            21       JAVA, so JAVA people typically write code in the 
            22       JAVA language.  That is compiled or, quote, 
            23       compiled into something called JAVA bite code.  
            24       JAVA bite code is a lower level language than 



             1       JAVA source code.  However, it's designed to be 
             2       machine independent.  When people ship around 
             3       little JAVA applets, it's the JAVA bite code 
             4       that's getting shipped around the network. 
             5                  Then when you get that to the 
             6       recipient machine, the bite code could either be 
             7       translated on the fly, which is a process called 
             8       interpretation, or for other purposes it itself 
             9       can be compiled by a compiler that's specific for 
            10       the target machine.  So that's an example where 
            11       you have three levels.  And in that case 
            12       something like JAVA bite code would be called an 
            13       intermediate language.
            14   Q.  So you would have source code, intermediate 
            15       language and then object code?
            16   A.  Then object code. 
            17   Q.  Now --
            18   A.  Although I should say it depends on who's 
            19       looking.  So you'll often find JAVA bite code 
            20       described as object code.  It really is kind of 
            21       in the eye of the beholder. 
            22   Q.  When you say it's "in the eye of the beholder," 
            23       then, you also talked about object code showing 
            24       you operations that are critical to how the code 



             1       operates, how the machine operates.  Would you 
             2       call that expression?  Would you explain that? 
             3   A.  Sure, I'll give you a --
             4                  MR. HART:  I'm going to object to 
             5       that, but go ahead. 
             6   Q.  Let me rephrase the question so there's no 
             7       problem about it --
             8   A.  Okay.
             9   Q.  -- in case we --
            10                  You used the term "operations 
            11       critical" in response to Mr. Hart's question.  
            12       Can you expand on that?
            13   A.  I'll give you an example.  So often an algorithm 
            14       will have a particular interlude which you want 
            15       to make run absolutely as fast as possible.  One 
            16       of the typical ways to make things run fast is to 
            17       take advantage of a property of the machines 
            18       called pipeline, which effectively means they can 
            19       set up to do the next operation while they're 
            20       doing the first one.  And often you can have as 
            21       many as, oh, ten different machine instructions 
            22       that are all in very different phases of 
            23       execution, which means your algorithm will run, 
            24       say, ten times as fast. 



             1                  Now, whether or not you're actually 
             2       able to do that depends on very subtle details of 
             3       how the algorithm is designed and subtle details 
             4       of the compilation process. 
             5                  So a particular -- just to give you a 
             6       particular example, if I have to fetch three 
             7       numbers from memory and do two multiplications, 
             8       the particular order in which you fetch things 
             9       from memory and do the multiplications will have 
            10       an impact on how well those things can be 
            11       pipelined. 
            12   Q.  And --
            13   A.  So I was going to say when somebody is designing 
            14       the interlude of an algorithm, it's very, very 
            15       common that you will in fact write a piece of 
            16       this algorithm in source code, run it through the 
            17       compiler to look at what the computer -- look at 
            18       the object code that the compiler produces, and 
            19       then modify that object code and feed that into 
            20       your, feed that into your program.  That's an 
            21       example of what I was talking about before when 
            22       you have an algorithm that is actually input to 
            23       the computer by a human as a mixture of source 
            24       and object code.



             1   Q.  Now, before Mr. Hart asked you about that 
             2       CitiBank analogy.  Can you give me the basis for 
             3       your statement or your view that you would give 
             4       CitiBank a certain amount of time; and if they 
             5       did not respond, then release the information 
             6       publicly?
             7   A.  I think it's -- I mean, I don't quite know what 
             8       it looks like to be sitting at CitiBank and 
             9       someone calls you up and says there's a major 
            10       problem in your banking system.  I don't know how 
            11       much you pay attention to that or how much you 
            12       would even consider it.  I don't know that if you 
            13       were an employee at CitiBank who got such a call, 
            14       perhaps in the programming department, and you 
            15       have to convince some vice-president to make a 
            16       major investment in changing their banking 
            17       system, I don't know how much luck you would have 
            18       doing that unless there is some real visible 
            19       impact to them taking no action.
            20   Q.  What is the ethical or social value of you 
            21       releasing that information five days, ten days, 
            22       or whatever, if a CitiBank doesn't respond to 
            23       you?
            24                  MR. HART:  Object, form, prior 



             1       testimony. 
             2   Q.  Go ahead. 
             3   A.  Say again. 
             4   Q.  Why would you do it?
             5   A.  Why would I give them five days notice, or why 
             6       would I give them only five days notice?
             7   Q.  Why would you give them notice, and why would 
             8       you, if they did not respond, then release the 
             9       information?
            10   A.  Well, I would give them notice because I think 
            11       it's the polite thing to do.  I mean, if my goal 
            12       is to actually make things, actually improve 
            13       things as opposed to my goal being to rob 
            14       CitiBank, then for one thing it demonstrates 
            15       intent.  This is my goal, I've done it publicly, 
            16       it demonstrates my good will.  Hopefully someone 
            17       now at CitiBank has seen this; if they know this 
            18       is going to happen, they would get some extra 
            19       warning.
            20   Q.  And if they do nothing, there's a possibility, is 
            21       there a possibility that people will then go in 
            22       and rob CitiBank on the basis of the information 
            23       you've given them? 
            24   A.  Yes, but offhand it's -- in general if you find a 



             1       flaw in a security system, it's wildly unlikely 
             2       that you're the only person who has found that 
             3       flaw.  And I think the real danger is that there 
             4       are people who are trying in fact to rob CitiBank 
             5       and are looking at this stuff very, very 
             6       carefully.  I think by the time this stuff gets 
             7       out to the public, they're unlikely to be the 
             8       first to know.
             9   Q.  So is it your testimony that you're being 
            10       protective, in a way, of CitiBank by releasing 
            11       the information?
            12   A.  I think I'm doing them a service.  I'm not 
            13       exactly sure -- I'm not sure they would 
            14       characterize it as protective of them.
            15   Q.  Why do you think you're doing them a service?  
            16       Why do you think you're doing them a service? 
            17   A.  Because I'm strongly encouraging them to improve 
            18       their system, with the likelihood that they are 
            19       guarding against security attacks that they do 
            20       not know about.
            21   Q.  Now, you say the DeCSS shows you how weak CSS is?
            22                  MR. HART:  Object, form and prior 
            23       testimony.
            24   Q.  Can you expand on that?



             1                  MR. HART:  Same objections.
             2   A.  Well, the descriptions -- be careful.  The 
             3       descriptions I have read of DeCSS, based on other 
             4       people who have seen it, demonstrate that it is 
             5       extremely weak.
             6   Q.  Do you know how and do you know where?
             7   A.  Two primary ways:  Apparently it's 40-bit 
             8       encryption.  So given merely that you know it's 
             9       40-bit encryption, I would expect you could 
            10       break, you could break something in a couple of 
            11       minutes. 
            12                  And the second way it's weak is that a 
            13       DeCSS recorder holds a decryption key -- I'm 
            14       sorry, DeCSS player holds a decryption key that 
            15       will decrypt any CD, so it's roughly -- it's 
            16       roughly as if, if I understand it correctly, 
            17       every recorder holds -- if I have a box that I 
            18       give everybody that has a key to your house, and 
            19       this key will open the lock on everybody's house, 
            20       so all somebody has to do is compromise one 
            21       recorder, and then you can decrypt every DVD.
            22   Q.  Do you know that in fact those recorders have 
            23       been compromised?
            24   A.  No, I don't know. 



             1   Q.  Do you know when the first cracks of CSS were 
             2       published?
             3   A.  No, I don't know. 
             4   Q.  You mentioned before, you used the term 
             5       "theoretical capacity" in response to Mr. Hart's 
             6       question.  What are the variables that affect 
             7       theoretical capacity?
             8   A.  That affect the theoretical capacity? 
             9   Q.  Yeah.  Concerning the computations you gave 
            10       before. 
            11   A.  I want to make sure I understand the question. 
            12   Q.  Go ahead.
            13   A.  You're asking if I were looking at a network and 
            14       saying what is it's theoretical capacity? 
            15   Q.  Yes.  And what are the characteristics that you 
            16       consider when you get to its real capacity?  In 
            17       other words, how is the theoretical capacity --
            18   A.  Okay.  Those are two different questions. 
            19   Q.  Pardon me. 
            20   A.  So theoretical capacity has to do with sort of 
            21       the basic bandwidth of the individual pieces of 
            22       the network, as I might measure in the 
            23       laboratory.  It might have to do with, if I took 
            24       one of the switches and drove it under ideal 



             1       circumstances, how fast it could switch. 
             2                  Let me think what else. 
             3                  Generally it doesn't consider 
             4       bottlenecks coming from the individual servers at 
             5       the end of the system.
             6   Q.  Can you describe to me what you mean by 
             7       bottlenecks from the servers at the end of the 
             8       system?
             9   A.  Quite simply, if my machine is trying to serve up 
            10       a file and trying to do 50 other things at the 
            11       same time, it may not be able to feed the network 
            12       as fast as it possibly could.  Similarly with the 
            13       machine on the other side that's receiving. 
            14                  Similarly, there are addressing 
            15       bottlenecks in the Internet.  If I have to go 
            16       through -- if I'm connected through something to 
            17       something by multiple hops, and depending on how 
            18       those switches are configured, there can be 
            19       slowdowns there.
            20   Q.  Is distance a factor?
            21   A.  Well, distance in --
            22                  MR. HART:  Object, form. 
            23   A.  -- itself is not a factor.  Distance in the sense 
            24       of the number of hops that you have to go through 



             1       is a major factor.
             2   Q.  Can you tell me why and how?
             3   A.  Because every time -- what happens is sort of a 
             4       piece of information goes over one segment of 
             5       network, and then gets effectively to a switch or 
             6       sometimes to a machine that acts as an 
             7       intermediate way station for it.  And then the 
             8       rest of the path has to be recomputed.  And that 
             9       takes a little bit of time.  By the time you get 
            10       a couple of hops, those things add up.
            11                  MR. GARBUS:  Now, Bill, I presume we 
            12       don't have to go through the affidavit line by 
            13       line?  I presume we can stipulate if he were 
            14       asked questions with regard to each allegation in 
            15       the affidavit, he would confirm it as his 
            16       testimony.  If you want, I'll show it to him and 
            17       say, "Is that your testimony?"
            18                  MR. HART:  I think I understand what 
            19       you're asking me, and without waiving any 
            20       objection to the fact that you're trying to 
            21       perpetuate the testimony, if your question is, 
            22       would Dr. Abelson testify to the effect that he 
            23       set forth in his affidavit, I think you can just 
            24       ask that question and have him answer it.



             1       BY MR. GARBUS: 
             2   Q.  Would you, if I were to ask you questions about 
             3       your affidavit, testify that the allegations in 
             4       your affidavit are accurate and true today?
             5   A.  Yes. 
             6   Q.  In the Bernstein case, do you know if it was 
             7       object, source code, or a combination?
             8   A.  I believe that Bernstein himself published source 
             9       code, and also I believe technical memos that 
            10       weren't encoded at all.  And I believe the 
            11       government's injunction covered both of those. 
            12                  There may have been arguments before 
            13       the Court about source and object code, but I 
            14       didn't look at those.  I don't believe there was 
            15       a distinction --
            16   Q.  You don't --
            17                  MR. HART:  I'm sorry, let the witness 
            18       complete his answer.
            19   A.  I don't believe there were distinctions made, but 
            20       I don't remember.
            21   Q.  In Younger, do you know if it was object code or 
            22       source code?
            23   A.  No, I don't.  Under the impression --
            24   Q.  Put aside what the Court said.  Do you know what 



             1       it was, in fact? 
             2                  MR. HART:  Object.
             3   A.  I don't quite remember exactly what the Court 
             4       said, but I don't believe people were talking 
             5       about those distinctions very much.
             6   Q.  Why? 
             7   A.  Because they were just talking about, again --
             8                  MR. HART:  Continuing objection to 
             9       this.  Go ahead.
            10   A.  Again, I don't remember exactly what Younger was 
            11       doing, but he was talking about teaching stuff in 
            12       his class, and I believe it was just discussing 
            13       things at the algorithm level, but I'm not sure. 
            14   Q.  Insofar as communicative expression, then, to 
            15       your mind, or to your, based on your education 
            16       and writing and teaching, are source code and 
            17       object code both covered as communicative 
            18       expression?
            19                  MR. HART:  Object, form. 
            20   A.  Well, both -- yes.  I mean, both -- if I take a 
            21       particular program, and I look at it in its 
            22       source form or in its object form, those express, 
            23       those communicate, those communicate different 
            24       things, and depending on what I want to do, I'll 



             1       want to look at them in different forms.
             2   Q.  Does the fact that you haven't actually seen the 
             3       DeCSS source code or object code change your 
             4       opinion that DeCSS object and source code are 
             5       both covered as communicative expression?
             6                  MR. HART:  Same objections and prior 
             7       testimony.
             8   A.  Change it from?  I haven't looked at it so....
             9   Q.  Instead of change it from, is it your view that 
            10       even though you have not looked at it thus far, 
            11       that DeCSS source and object code both come 
            12       within communicative expression?
            13                  MR. HART:  Same objections.
            14   A.  I would expect they do. 
            15   Q.  Why? 
            16   A.  As I said, I assume that someone inferred that 
            17       there's a 40-bit key from looking at the source 
            18       code, where the object is, as an example of 
            19       something expressed.  Someone inferred that 
            20       there's a global master key that is sitting on a 
            21       DeCSS recorder, from looking at either the source 
            22       or object code.  So that certainly would count as 
            23       something that was expressed. 
            24                  I would imagine that if I were -- if 



             1       someone wanted me to design a, say, a computer 
             2       program -- in fact, I don't know.  I don't know 
             3       how the various player programs work, but I 
             4       assume that if someone asked me to write 
             5       something that simultaneously decrypted something 
             6       while it played it, or if someone said to me, 
             7       "Try to minimize the amount of intermediate 
             8       storage that a software program that would play a 
             9       DVD would work," that's a place where I would 
            10       want to look at the object code, specifically to 
            11       talk about these pipelining issues that I 
            12       mentioned before. 
            13   Q.  Now, Mr. Hart asked you about your schedule.  The 
            14       trial starts July 17.  Tell me your schedule for 
            15       the 17th.
            16   A.  What day of the week is the 17th?
            17   Q.  A Monday. 
            18   A.  I am planning to be on the West Coast right now 
            19       on the 17th.  What I can say pretty, with pretty 
            20       much certainty, is that I will be on the East 
            21       Coast either the week of the 17th or the week of 
            22       the 26th, and I haven't decided which one yet.
            23   Q.  And is that consistent with your Hewlett Packard 
            24       obligations?



             1   A.  Yes.
             2   Q.  That's what draws you out there?
             3   A.  That's what draws me out there. 
             4   Q.  Now, Mr. Hart asked you some information about 
             5       transmission times.  Let me show you an affidavit 
             6       of Michael Shamos. 
             7                  MR. GARBUS:  Can we mark this as an 
             8       exhibit?  I only have one copy. 
             9                  By the way, you do not have DeCSS or 
            10       CSS with you, or the object or source code.
            11                  MR. HART:  I think you asked me that 
            12       before, and I think I told you no.
            13                  MR. GARBUS:  Okay. 
            14                  (Document marked as Exhibit 5
            15                  for identification.)
            16                  MR. HART:  May I see what you're 
            17       showing the witness, please? 
            18                  MR. GARBUS:  Let the record indicate 
            19       I'm showing the witness an affidavit from a 
            20       Martin Shamos, it's the identical affidavit.  The 
            21       only thing that's different is on the front page 
            22       in my handwriting is Mr. Abelson's phone number.
            23                  MR. HART:  There's also a handwritten 
            24       notation on the 6th page of this.



             1                  MR. GARBUS:  Yes, there is. 
             2                  MR. HART:  You know, without 
             3       belaboring, this I sort of object to your handing 
             4       the witness something that somebody else has 
             5       marked on.  That's not a true copy of the 
             6       exhibit. 
             7                  MR. GARBUS:  Okay.
             8                  MR. HART:  If I said "I sort of 
             9       object," I strenuously object. 
            10                  (Document exhibited to witness.)
            11   Q.  I show you --
            12   A.  Is there something I'm not supposed to see here? 
            13   Q.  No, no, no. 
            14                  I show you the affidavit of Michael 
            15       Shamos, and I ask you, with respect, if you can, 
            16       with respect to his computations and his 
            17       analysis, if you can give me your views of that. 
            18   A.  So I've seen this before.  You faxed it to me a 
            19       couple of days ago.  That's my fax number that's 
            20       written on there. 
            21                  I've been thinking about it some.  I 
            22       think the -- I'm certain the facts in there look 
            23       true.  Shamos has a very good reputation.  I 
            24       think he's very well respected.  We were both at 



             1       Princeton at the same time.  He had a very good 
             2       reputation at Princeton. 
             3                  There's nothing in here to disbelieve 
             4       his, except that it's based on -- the 
             5       implications are based on theoretical capacity.  
             6       So I notice he talks about transmitting a DVD 
             7       movie in less than 20 minutes, and I think -- so 
             8       it looks like I got the test you gave me right, I 
             9       said about 15.  That sounds about right for 
            10       transferring it. 
            11                  The one thing that he doesn't take 
            12       account of, though, is actual network capacity as 
            13       opposed to bandwidth.  So if you go back to -- 
            14       you asked me about Napster before, and university 
            15       banning it. 
            16                  MIT, as I said before, didn't ban it 
            17       or didn't have a problem with it, but many 
            18       universities did, and they had a problem because 
            19       of capacity.  The notion here that a lot of 
            20       people would be transmitting DVDs is going to, is 
            21       probably going to be impossible because of 
            22       constraints on network capacity. 
            23                  So, for example, let's suppose I take 
            24       a DivX compressed DVD.  If I use Shamos's numbers 



             1       here -- what does he say?  He doesn't say 
             2       exactly.  But assume it's, when I compress it, 
             3       it's 600 megabytes, so it's roughly a CD.  And 
             4       then he talks about --
             5   Q.  Excuse me.  Can we just --
             6                  MR. HART:  I'd like the witness to 
             7       complete his testimony rather than you interrupt 
             8       him.
             9   Q.  I know that.  Can we use as an assumption that a 
            10       DVD is somewhere between 5 and 8 gigabytes?
            11                  MR. HART:  Objection.  You're 
            12       testifying.
            13                  THE WITNESS:  Wait, wait. 
            14                  MR. HART:  The witness was saying 
            15       something different, and you have interrupted 
            16       him.
            17                  Would you mark this, please, Ms. 
            18       Reporter, because I think that was inappropriate, 
            19       Marty.
            20   A.  Let me speak to, actually --
            21   Q.  Okay.  Do it without my reference. 
            22   A.  What Shamos did where -- what he's talking about 
            23       is pushing around the compressed form, which is 
            24       what he got from this other person over the 



             1       Internet, in 20 minutes.  Let's leave aside what 
             2       an uncompressed, what he says here, 
             3       multigigabyte.  Let's only talk about the 
             4       compressed one, the thing he's actually writing 
             5       to. 
             6                  So he transmitted -- what did he do?  
             7       He transmitted that from this other person in, he 
             8       said, 11 hours, which sounds reasonable, given 
             9       that that person was probably at home, or 
            10       something.  Then he said that if that were going 
            11       on between students in the university, you could 
            12       do the same thing in 20 minutes.  That's true 
            13       based on bandwidth. 
            14                  However, if you assume that that's a 
            15       600 megabyte file, then transmitting a 600 
            16       megabyte file in 20 minutes is a data rate of 
            17       something like 4 megabits per second.  So before 
            18       we were talking about 10 or a hundred.  So if you 
            19       assume -- assume you're only using 4 megabits per 
            20       second. 
            21                  Now, at a place like MIT, you have a 
            22       hundred megabits per second available for 
            23       everybody.  So what that says is that if 25 
            24       people were simultaneously doing this, you would 



             1       completely saturate the MIT network.  And I can 
             2       assure you if five people were simultaneously 
             3       doing this, this would have a pretty big impact 
             4       on the MIT network.
             5   Q.  Why would it saturate the network?
             6   A.  Because -- when you say that the network is a 
             7       hundred megabits per second, so at MIT that is a 
             8       shared hundred mega bits for everybody, and that 
             9       means that if people are using -- if there are 25 
            10       people, each of whom are transferring at 4 
            11       megabits per second, then you've used up the 
            12       entire network capacity for anything at all, 
            13       electronic mail, messaging, anything. 
            14                  So I think what you find is the same 
            15       phenomenon that encouraged some universities to 
            16       close down Napster; you'd have that phenomenon 
            17       multiplied by ten or a hundred using these 
            18       numbers.  And certainly well before you got to a 
            19       lot of DivX trading, you would have universities 
            20       shut this down, quite apart from any issue about 
            21       liability or legality or anything. 
            22                  The same phenomenon -- he goes on and 
            23       talks about Internet 2.  So my understanding of 
            24       Internet 2 right now is that there's about 2.4 



             1       gigabits per second.  You've got to remember, 2.4 
             2       gigabits per second shared for the whole country.  
             3       So that means if you had 6,000 people who were 
             4       simultaneously pushing one of these DivX files 
             5       around, you would shut down all of Internet 2 for 
             6       all university connections for the whole country. 
             7                  So the analogy is, imagine that there 
             8       was a special kind of phone call that somebody 
             9       could make, and this phone call took 20 minutes.  
            10       And if at any given time you had 6,000 people 
            11       somewhere making, somewhere making this 20-minute 
            12       phone call, you shut down the entire U.S. phone 
            13       system.  It just wouldn't work. 
            14                  I don't know what the actual numbers 
            15       can be, but this, but the notion that simply 
            16       because the theoretical bandwidth is there that 
            17       you would actually have that kind of trading 
            18       going on on the Internet is a little hard to 
            19       believe, certainly in this generation of Internet 
            20       technology that's available to universities, 
            21       almost certainly in the next one, which is 
            22       Internet 2, and you can't quite say after that. 
            23   Q.  When you say this generation, the next 
            24       generation, how many years are we talking about?



             1   A.  Internet 2 is just starting.  And what are people 
             2       talking about?  MIT is just hooking up to it.  I 
             3       would expect -- I don't quite know the roll-out 
             4       plans.  I would expect it's something like two 
             5       years.  I haven't even heard people talking very 
             6       much about generations after that. 
             7   Q.  Now, when you talk about MIT's capacity, is that 
             8       larger than most universities, smaller than most?  
             9       Can you give me a context?
            10   A.  It's larger than most. 
            11   Q.  Can you tell me something about that? 
            12   A.  It has to do with the whole infrastructure.  So, 
            13       again, you can lay cables -- it has to do with 
            14       how many routers and how many switches you have 
            15       and how well you keep them up to date. 
            16                  I don't have any facts on 
            17       universities, but, again, as I said, Napster 
            18       wasn't a particular problem for MIT, and I know 
            19       that it was for lots of other universities.  But 
            20       whether that's a factor of 2 bit or a factor of 4 
            21       bit or a factor of 10 bit, I don't know.
            22   Q.  You've talked about transferring music files, 
            23       Napster, and transferring movie files, let's say 
            24       through DVD.  What is the difference in size of 



             1       the material being transferred and the amount of 
             2       time that -- the time difference?
             3                  MR. HART:  Object, form. 
             4   A.  Well, the difference is the size of the files you 
             5       have to transfer.
             6   Q.  Can you give me some sense of that difference?
             7   A.  Well, let's use Shamos' numbers.  There are three 
             8       stages.  There's uncompressed video, like you 
             9       would get out of running DeCSS.  So we use 
            10       Shamos's numbers, he says those are 
            11       multigigabytes, so, I don't know, say 6 
            12       gigabytes.  Then there's compressed down to DivX, 
            13       which I gather from his numbers is something like 
            14       600 megabytes.  But, again, I don't know what 
            15       quality that is.  I believe they compress to 600 
            16       megabytes because people want to write them on 
            17       CDs, but I don't know what quality compromise you 
            18       make from going to multigigabytes to 600 
            19       megabytes.  Let's use the 600 megabytes. 
            20                  600 megabytes is about what, is about 
            21       what fits on an audio CD.  So let's assume that's 
            22       an hour of uncompressed audio music. 
            23                  Now, if we compare that with Napster, 
            24       Napster, first of all, is transmitting MP3, which 



             1       is about a factor of 10 over, factor of 10 
             2       compression.  So you start with your one CD and 
             3       you get a factor of 10 for compression.  And how 
             4       long is a song, five minutes?  So that's a 20th 
             5       of an hour.  You tell me.  I'm thinking about 
             6       people trading Napster songs.  Let's say it's a 
             7       10-minute song.  So that says it's transmitting 
             8       one DivX, one DivX compressed feature length 
             9       movie is a hundred times transmitting a Napster 
            10       song.
            11   Q.  If it's not compressed?
            12                  MR. HART:  Objection, form.
            13   A.  If the movie is not compressed.
            14   Q.  Then what is the --
            15   A.  God.  So we're doing 600 megabytes to -- let's 
            16       use 6 gigabytes.  So it's another factor of 10.
            17   Q.  So what is that?
            18   A.  A thousand Napster songs. 
            19   Q.  Now, you said -- you talked about Internet 2, 
            20       let's say, having a life of two years until 
            21       something --
            22   A.  I believe they're rolling it in over two years.  
            23       But, again, I'm not sure.
            24   Q.  And tell me how you arrive at that?  How do you 



             1       know that?
             2   A.  This is just impressions from talking to people.  
             3       Again, MIT has been planning for this for a year, 
             4       and we don't quite have it yet.  And we tend to 
             5       be closer to the backbone than a lot of other 
             6       universities.
             7   Q.  Do you know of any other universities that have 
             8       gone into Internet 2 yet?
             9   A.  Probably -- Shamos mentioned CMU.  They're 
            10       certainly very early in putting the technology 
            11       on.  I would expect Stanford, I would expect 
            12       Cornell, UCLA.
            13   Q.  Now, with respect to the consumer -- go ahead.
            14   A.  No, I'm just trying to....
            15   Q.  When you talk about Internet 2, you're talking 
            16       about something that the universities are 
            17       starting to deal with.  With respect to the 
            18       consumer, how long is it before you will have 
            19       this next generation available? 
            20                  MR. HART:  Objection to form. 
            21   A.  I don't --
            22                  MR. HART:  And prior testimony. 
            23   A.  I don't know.  I don't know of plans that are 
            24       much beyond cable modems going to houses.  



             1       There's talk about people trying to, but I know 
             2       of no plans that are particularly faster than 
             3       what you get over cable modem.
             4   Q.  I draw your attention in the Shamos affidavit 
             5       where he says he uses DeCCS software.  What is 
             6       the significance or lack of significance of him 
             7       using DeCCS to do his experiment in transmission?  
             8                  MR. HART:  Objection, form, prior 
             9       testimony. 
            10   A.  Well, that's one of the things he doesn't, that 
            11       doesn't quite come out here.  One of the things 
            12       that you don't know is he went through this whole 
            13       process, which started with a DVD and ended up 
            14       with a compressed DivX file.  One step in that 
            15       process is running DeCSS.  And then the question 
            16       is in terms of the overall process that you have 
            17       to go through, how much of that is DeCSS; and if 
            18       you didn't have DeCSS, could you effectively get 
            19       the same thing. 
            20                  I would imagine he could have just 
            21       taken his -- what did he do -- a video player, he 
            22       could have just taken the output of whatever he 
            23       played it on and intercepted the signal and 
            24       gotten a computer file, and he would have had 



             1       something that is functionally equivalent to 
             2       having run DeCSS on that thing. 
             3                  And then by the time you take that and 
             4       you compress it through DivX, I would be actually 
             5       pretty surprised if you could see the difference 
             6       between someone starting with a DVD, running 
             7       DeCSS to get the output, and compressing it, 
             8       versus starting with the DVD, just grabbing the 
             9       output of the video player and compressing it and 
            10       producing the same file.  And one of the things 
            11       -- I wouldn't even know which process would be 
            12       more difficult.
            13   Q.  When you say "more difficult," you mean also 
            14       faster and more efficient?
            15   A.  Faster and more difficult.
            16   Q.  You don't know which one would produce a better 
            17       image?
            18   A.  I would -- again, I would venture to guess that 
            19       you couldn't tell the difference.  But someone 
            20       ought to --
            21                  MR. HART:  Wait.  Again, I'd like the 
            22       witness to finish.
            23   A.  I would venture to guess you couldn't tell the 
            24       difference, but someone ought to try that and 



             1       see. 
             2   Q.  Do you have any sense of what you lose when you 
             3       do a compression of the kind that has been 
             4       described in the loss of image?  Have you ever 
             5       seen anything with that degree of compression? 
             6                  MR. HART:  Object, form. 
             7   A.  No, I don't know.  There's also, there's the 
             8       sound compression that you have to worry about, 
             9       too. 
            10                  One of the things I did notice is that 
            11       there's this step in here where he talked about 
            12       having to fix the desynchronization.  For all I 
            13       know, if you grab the video from the output 
            14       recorder, you wouldn't even have to do that step.  
            15       But, again, I've never done either of those.
            16   Q.  Tell me why you think you might not have to do 
            17       that step?
            18   A.  Because you're getting the output synchronized 
            19       from the recorder.  If it comes out as a MP3 
            20       file, then you just have it synchronized.
            21   Q.  In a way, DeCSS is less user friendly than this 
            22       other process?
            23                  MR. HART:  Objection, form, prior 
            24       testimony. 



             1   A.  Well, I'm not sure that either of them is less -- 
             2       I'm not sure I would characterize either of them 
             3       as user friendly.  It's not clear that if I'm in 
             4       the business of wanting to copy videos to send 
             5       them over the network to a friend somewhere, it's 
             6       not clear to me having DeCSS makes a big 
             7       difference in whether or not I can do that. 
             8                  MR. GARBUS:  I have no further 
             9       questions. 
            10                  MR. HART:  I'm going to have a few, 
            11       but I would like to take five minutes before I 
            12       do. 
            13                  (Brief recess.)
            14                  MR. GARBUS:  I'd like to ask the 
            15       witness a few more questions that came to me 
            16       during the break. 
            17                  MR. HART:  While you were talking to 
            18       the witness off the record?
            19                  MR. GARBUS:  Oh.  Sure. 
            20       BY MR. GARBUS: 
            21   Q.  You had indicated before that your site linked to 
            22       the Touretzky site; is that correct?
            23   A.  Yes.
            24   Q.  And you indicated that the Touretzky site 



             1       contains various material.  To your knowledge, 
             2       does it contain source codes and object codes for 
             3       DeCSS?
             4                  MR. HART:  Object, asked and answered, 
             5       prior testimony.
             6   A.  I believe it had links to a whole bunch of 
             7       different descriptions of DeCSS.  I think some of 
             8       them -- I think one was source code.  I think one 
             9       was a picture.  I think one was a T-shirt.  I 
            10       didn't look in detail.  I don't know whether it 
            11       had object code or not.
            12   Q.  Is it your perception that you were doing 
            13       something illegal in linking to a site that might 
            14       have object or source code --
            15                  MR. HART:  Object.
            16   Q.  -- of DeCSS?
            17                  MR. HART:  Same objections, prior 
            18       testimony.
            19   A.  No, of course -- it's certainly not illegal.
            20   Q.  Do you know what the World Wide Web Consortium 
            21       was or is?
            22   A.  Yes.  I was part of setting up the World Wide Web 
            23       Consortium at MIT.
            24   Q.  And what is it?



             1   A.  It's an organization whose mission is, quote, "to 
             2       bring the Web to its full potential," and it does 
             3       this by advocating different kinds of standards 
             4       and interoperability standards.
             5   Q.  Is the consortium just a private organization or 
             6       is the government involved in it?
             7   A.  It's a group of mostly members who are private 
             8       companies.  I don't remember whether there are 
             9       any government organizations who are members, but 
            10       there may be.  And it's sponsored by a couple of 
            11       universities, MIT being one of them.
            12   Q.  Who are some of the private companies that are 
            13       involved, if you know?
            14   A.  Microsoft, Hewlett Packard.  I believe the Direct 
            15       Marketing Association is a member.  I forget.  I 
            16       think there are about 400 members.
            17   Q.  And have they expressed any views on linking?
            18                  MR. HART:  Object, the obvious 
            19       reasons. 
            20   A.  The World Wide Web Consortium takes the position 
            21       that linking is a reference, and it's fundamental 
            22       to the architecture of the Web that people link 
            23       and that you have these references, and that's 
            24       how the Web is glued together.



             1   Q.  Does the name Timothy Berners-Lee mean anything 
             2       to you?
             3   A.  Tim Berners-Lee, he's the chairman of the Web 
             4       Consortium and the inventor of the World Wide 
             5       Web.
             6   Q.  And the author of a book called Weaving the Web?
             7   A.  Yes.
             8   Q.  Do you know on the Web you can link to 
             9       information about making bombs?
            10   A.  I believe you can.  You can discuss that it 
            11       exists.
            12   Q.  And you can link to sites that give you the 
            13       formula for seran gas?
            14   A.  I believe you can. 
            15   Q.  Do you know what seran gas is?
            16   A.  I guess it's a poison gas, very dangerous poison 
            17       gas.
            18   Q.  And is it your view, as someone who is a member 
            19       of the world wide consortium, that linking should 
            20       be permitted to things like DeCSS or bomb making 
            21       or the creation of seran gas?
            22                  MR. HART:  I object.
            23   Q.  And can you break it down with respect to any one 
            24       of those three?



             1                  MR. HART:  I object, and also on 
             2       competence grounds.  Are you testifying on behalf 
             3       of the world wide consortium?
             4                  MR. GARBUS:  No,  he's not.
             5                  MR. HART:  I maintain my objection. 
             6                  MR. GARBUS:  Go ahead. 
             7   A.  In my view, telling people that information is 
             8       available in general should not be prohibited.  I 
             9       think that by -- you're not implying advocacy or 
            10       responsibility or anything, you're just telling 
            11       someone that something is available somewhere. 
            12   Q.  Now, when was the World Wide Web formed?
            13   A.  The Web itself was invented, I believe, in 1990, 
            14       I think.  And it came to MIT in 1995, 1994.  I 
            15       believe that's when the consortium was set up.
            16   Q.  And the consortium exists today?
            17   A.  Yes.
            18   Q.  And how many academics are part of that 
            19       consortium or members of the consortium?
            20   A.  Individual academics? 
            21   Q.  Yes. 
            22   A.  Formally, organizations join.  Most of the 
            23       organizations are companies.  But many people who 
            24       are informally associated with it and their 



             1       conferences, generally, are in large part 
             2       academics who are writing things about Web 
             3       standards and Web technology.
             4   Q.  And have you done any writing in that area?
             5   A.  Not in any of the Web Consortium journals.
             6   Q.  Elsewhere?
             7   A.  Mostly been informally involved with working with 
             8       people in the Consortium design standards.
             9   Q.  Do you, in any of your classes, or, to your 
            10       knowledge, do any of the MIT professors teach 
            11       object code or how to read object code in their 
            12       classes?
            13   A.  Oh, I -- teach how to read -- there certainly are 
            14       many classes in which you read object code.
            15   Q.  Is that something that's taught also, how to read 
            16       it?
            17   A.  Yeah, how to read it, how to write it.  I would 
            18       have said that up until ten years ago, it was 
            19       completely standard to program, to have courses 
            20       in programming what's called ASSEMBLY code.
            21   Q.  Is it your experience that students now come into 
            22       MIT, many of them already knowing how to write 
            23       object code?
            24   A.  Many?  Yeah, I think "many" is correct.  Again, 



             1       it depends on the particular machine.
             2   Q.  How long have you been teaching at MIT?
             3   A.  Since 1973.
             4   Q.  And what is your status there?
             5   A.  Full professor.
             6   Q.  How long have you been a full professor?
             7   A.  Good question.  I don't remember.  1983, perhaps.  
             8       1984.
             9   Q.  Where were you educated at?
            10   A.  I have a graduate degree from MIT and 
            11       undergraduate degree from Princeton. 
            12   Q.  And the curriculum vitae that has been marked as 
            13       Exhibit 1, is there anything you would add to it 
            14       or is there anything more recent or anything more 
            15       relevant to today's discussion that you can think 
            16       of?  If not....
            17   A.  You might want to add the -- I mentioned the 
            18       joint course I teach with Harvard Law School, but 
            19       other than that....
            20                  MR. GARBUS:  Go ahead, Mr. Hart.  I'm 
            21       through.
            22                  MR. HART:  You're really done?  I 
            23       maintain my objection to all of that.
            24                  MR. GARBUS:  Unless I have another 



             1       thought. 
             2                  MR. HART:  I do have a few questions 
             3       for you.
             4                  REDIRECT EXAMINATION
             5       BY MR. HART: 
             6   Q.  You mentioned earlier that one of your areas of 
             7       consulting for Hewlett Packard was -- correct me 
             8       if I'm wrong -- to develop systems that were 
             9       password protected; is that right? 
            10   A.  I'm not sure that's the right question. 
            11   Q.  What is the right question? 
            12   A.  You mean different kinds of authorization 
            13       mechanisms for having people get at things.  Some 
            14       of them are passwords, some of them -- there's 
            15       work being done in smart keys, there's work being 
            16       done in biometrics.
            17   Q.  I didn't mean to confine it to passwords per se. 
            18                  What's the reason for having systems 
            19       that require an authorization before a person can 
            20       access the information that's on the system?
            21   A.  The deep reason is that you don't want to give 
            22       everybody access to everything you have.  The 
            23       proximate reason is that as organizations connect 
            24       more to the World Wide Web, they're opening up 



             1       their internal systems to a degree that was just, 
             2       no one contemplated before.  So as soon as you 
             3       have people potentially being able to access 
             4       anything on any of your computers, you want to 
             5       restrict access to all sorts of people within and 
             6       without the organization.
             7   Q.  Gotcha.  But, for example, there is a utility in 
             8       having such systems for the purpose of exchanging 
             9       scientific and research information; is that 
            10       correct?
            11   A.  Having --
            12   Q.  I said such systems, namely systems that require 
            13       an authorization code to enter. 
            14   A.  There's not a special one for scientific 
            15       information.
            16   Q.  I'm sorry, it was a general question. 
            17   A.  Generally scientific and technical information, 
            18       well, from a university would go out free, but of 
            19       course in an industry, I mean, much of scientific 
            20       and technical information would be confidential. 
            21   Q.  But research and scientific information can be 
            22       shared through such networks requiring an 
            23       authorization code, whether or not there's a 
            24       charge made for access; isn't that correct?



             1   A.  Oh, I wasn't thinking about consumer stuff at 
             2       all.  A more typical arrangement would be two 
             3       companies doing joint research.  They'll set up a 
             4       system that some group of people in company A can 
             5       share information with some group of people in 
             6       company B, and that's a very important mode of 
             7       collaboration because they might be working on 
             8       something that's confidential to the two of them. 
             9   Q.  And would that include activities like reverse 
            10       engineering?
            11   A.  Sure it could. 
            12   Q.  And is there any reason why that model would not 
            13       apply to any other kind of reverse engineering or 
            14       sharing of information in an engineering context, 
            15       namely authorization protected --
            16   A.  Wait a minute, I lost --
            17                  MR. GARBUS:  I don't understand the 
            18       question.
            19   A.  I lost where you're saying reverse engineering.
            20   Q.  Let me just try to simplify it.  You've talked 
            21       about the sharing of information over sort of a 
            22       private net through authorization and I'm --
            23   A.  Right.
            24   Q.  -- asking you whether that model has utility for 



             1       the sharing of scientific or engineering 
             2       information outside of strictly commercial 
             3       applications?
             4   A.  Yeah.  Yes, certainly. 
             5   Q.  Okay.  Now, during the course of the testimony 
             6       that you gave Mr. Garbus, you referred a couple 
             7       of times to DeCSS player or recorder.  I wanted 
             8       to make sure your testimony was clear in that 
             9       regard.
            10   A.  If I did, I shouldn't have said that, I should 
            11       have said DVD recorder or player. 
            12   Q.  Okay.  You also mentioned hops in the course of 
            13       your testimony.
            14   A.  Oh, okay. 
            15   Q.  Can you tell me what an average string or trace 
            16       of hops would be for a normal Internet 
            17       communication?  Is there an average? 
            18   A.  I don't know if there's an average.  Certainly 10 
            19       is pretty common.  I wouldn't know what an 
            20       average is.
            21   Q.  The more hops there are, the slower the transfer 
            22       rate or the potential risk that the transfer rate 
            23       is going to be slower?
            24   A.  In general. 



             1   Q.  And do you, the user, at one end of that chain 
             2       have any ability to control the number of hops by 
             3       how you route the signal? 
             4   A.  No, almost never, other than in sort of very 
             5       experimental settings by people doing network 
             6       research.  But the design of the Internet does 
             7       that routing adaptively on purpose.  So there are 
             8       a couple of experiments that people are working 
             9       on that allow you to control that, but an average 
            10       user has no access to that. 
            11   Q.  Do you?
            12   A.  No.  I mean, for example, the people who do 
            13       network research at the laboratory for computer 
            14       science where I work, for the experimental 
            15       networks they're working on, are able to control 
            16       that.
            17   Q.  Gotcha.  To improve network efficiency?
            18   A.  Right.
            19   Q.  And to reduce the number of hops?
            20   A.  Right.  That's one. 
            21   Q.  Now, you were talking with Mr. Garbus about the 
            22       consumption of bandwidth, if you will, when 
            23       multiple users are sharing files of different 
            24       sizes, right?



             1   A.  That's right.
             2   Q.  Is there a difference, to your understanding, 
             3       between a shared or hub-type network and a 
             4       switched network; and if so, what is that 
             5       difference?
             6                  MR. GARBUS:  That's already been asked 
             7       and answered, but go ahead. 
             8   A.  I don't -- I don't quite know how to characterize 
             9       that difference.  I'm not quite sure what you 
            10       mean. 
            11                  The particular estimates I was giving 
            12       were quite independent of that because they were 
            13       simply talking about what's the aggregate around 
            14       MIT as opposed to -- I think what you might mean 
            15       is there some master switch in which everything 
            16       at MIT has gone --
            17   Q.  No, no.  Let me try and help you. 
            18                  We were talking about bandwidth and 
            19       effective transfer rate as a function of 
            20       bandwidth, right? 
            21   A.  Right. 
            22   Q.  And you postulated that given the size of files, 
            23       even DivX files, that multiple users sending 
            24       multiple DivX files would consume available 



             1       bandwidth at a certain point, making the delivery 
             2       of movies over the Internet, at least now, 
             3       impractical; is that a fair --
             4   A.  Right.  What I said --
             5   Q.  -- statement of what you said in your direct 
             6       testimony a moment ago?
             7   A.  Yes.
             8                  MR. GARBUS:  Let the witness answer.
             9                  MR. HART:  I'm sorry, Mr. Garbus?
            10                  MR. GARBUS:  Go ahead, he's answered. 
            11   Q.  And my question is whether that testimony was 
            12       based on something called a shared network or a 
            13       hub network rather than a switched network; and 
            14       in that context, if there is a distinction 
            15       between those things, I would like you to tell me 
            16       what your understanding is of that distinction?
            17   A.  I don't know.  I don't think the estimates I was 
            18       making depend on that. 
            19   Q.  In any way whatsoever?
            20   A.  I think someone can make many more detailed 
            21       estimates that would reflect on that, but I was 
            22       saying an overall generalization.
            23   Q.  Right.  Again, not going to the particular 
            24       numbers in any case, number of minutes, but going 



             1       to the proposition that multiple users sharing 
             2       multiple files simultaneously would consume 
             3       bandwidth?
             4   A.  Yes.
             5   Q.  Going to that proposition and that proposition 
             6       only, my question is, is there -- do you have any 
             7       understanding of whether there's a difference in 
             8       regard to the use of shared networks and switched 
             9       networks as it affects consumption of bandwidth?
            10                  MR. GARBUS:  He's already answered.  
            11       Go ahead. 
            12   A.  Okay.  I think I understand.  This is something I 
            13       know about only vaguely.
            14   Q.  Only?
            15   A.  Only vaguely.
            16   Q.  Vehically?
            17   A.  Vaguely.  Vaguely. 
            18   Q.  I'm sorry. 
            19   A.  I think at the level that I was talking about, it 
            20       doesn't matter, because I was talking about the 
            21       network backbone at MIT, and so -- how do I say 
            22       this?  I don't know that you can set up a 
            23       switched connection at MIT, but assuming you 
            24       could, you're still drawing bandwidth from that 



             1       backbone.  So it doesn't matter whether you're 
             2       sharing it or switching it at that level. 
             3                  I know at MIT we have a hundred 
             4       megabits, and we can presumably choose to 
             5       allocate some of that by switching networks, 
             6       which I'm almost sure we do not, but it doesn't 
             7       make any difference for the aggregate numbers I 
             8       was talking about. 
             9   Q.  Is it your understanding that a switched network 
            10       enables multiple users to use the same amount of 
            11       bandwidth at the same time without consuming the 
            12       amount of bandwidth that's available to other 
            13       users on the same network?
            14   A.  I don't think that's quite the right way to 
            15       characterize it.  It has to do with the 
            16       granularity at which you're sharing things.  At 
            17       some point there's only so much bandwidth.  If I 
            18       switch among users rapidly enough, it may seem to 
            19       you that you're not sharing, but at the levels 
            20       where you start saturating the network, that 
            21       becomes absolutely apparent that the sharing is 
            22       going on. 
            23   Q.  Does the presence of multiple servers have any 
            24       impact upon your views in that regard?



             1   A.  Be a little more precise.  If I'm trying to get a 
             2       file and it exists at multiple sources, or simply 
             3       there's a lot of servers?
             4   Q.  Either one or both. 
             5   A.  Well, it will have an impact, but the overall 
             6       estimates remain the same. 
             7                  There are tricks you can do with 
             8       multiple servers to essentially cut down 
             9       bottlenecks and server delays.  But, again, the 
            10       kind of estimates I was talking about was 
            11       independent of that.  Simply talking about how 
            12       much the network itself can handle. 
            13                  I'll certainly grant you that if you 
            14       generalize from MIT to the national network, 
            15       there's a lot more complexity, but I don't think 
            16       it changes the fundamental estimates. 
            17   Q.  Now, is the type of network that you were 
            18       describing what one might call a hub-type 
            19       network?
            20   A.  I don't think so.
            21   Q.  What's the term for the type of network you were 
            22       describing?
            23   A.  At MIT, I believe MIT has what's called a ring.  
            24       So there's a central thing that lots of things in 



             1       the campus are connected to, and they all fit in 
             2       the ring.
             3   Q.  Are you aware if the Napster service depends upon 
             4       multiple servers?
             5   A.  The Napster service itself is just a directory.  
             6       I'm testifying beyond what I know, but my 
             7       assumption is they simply have a server 
             8       someplace.  I don't know if they have multiple 
             9       servers or not. 
            10   Q.  Now, the -- what's available to MIT students in 
            11       the dorm rooms, did you say it's 10 meg --
            12   A.  I believe --
            13   Q.  -- Ethernet?
            14   A.  I believe there's a mixture of 10 and a hundred.
            15   Q.  When you were talking about the value of DeCSS 
            16       versus not using DeCSS and you mentioned taking 
            17       the signal at the output of a player?
            18   A.  Right.  I would imagine you can do that.
            19   Q.  Are you talking about the analog output or 
            20       digital output or what?
            21   A.  Well, certainly you could do it with analog 
            22       output.  I would expect you could do it with 
            23       digital output.  But, again, I haven't done it. 
            24   Q.  Uh-huh.  Is there any difference in your 



             1       professional experience between the signal 
             2       quality of a digital signal versus an analog 
             3       signal, particularly when it's being replicated?
             4   A.  Well, the answer is that if you make multiple -- 
             5       digital copies generally are perfect, whereas 
             6       analog copies, you'll get distortion through 
             7       multiple copy generations. 
             8                  But remember, there's another -- if 
             9       this is in reference to what I was talking about 
            10       making with Shamos's thing, remember that you 
            11       still have to go through a step of compression, 
            12       and you are eventually producing a digital copy. 
            13                  If I take -- let's assume I make an 
            14       output, an analog output coming out of the 
            15       recorder.  So I take that analog output, I 
            16       digitize it, I compress it, and that whatever 
            17       degeneration I've gotten from that process, after 
            18       that I can ship those things around and they can 
            19       be replicated perfectly. 
            20                  So then you have to ask yourself, in 
            21       the overall degeneration between, of which most 
            22       will come from compression, I would guess, how 
            23       much extra quality loss you would get from going 
            24       from the recorder initially to analog as opposed 



             1       to doing a perfect copy from decryption, and I 
             2       was guessing that that wouldn't be noticeable.  
             3       But, again, I don't know.
             4   Q.  Gotcha.  You testified earlier in response to Mr. 
             5       Garbus's questions that you regarded notice to 
             6       the proprietor of a system that's been hacked or 
             7       breached as the polite thing to do, among other 
             8       things?
             9   A.  Right.
            10   Q.  And you said that it was indicia of good faith?
            11   A.  (Witness nods.)
            12   Q.  You have to verbalize your response.
            13   A.  Yes, that's what I said. 
            14   Q.  Does that mean that the absence of notice to the 
            15       proprietor would be, in your estimation, indicia 
            16       of bad faith?
            17   A.  Not necessarily.  It would just be the absence of 
            18       indicia of good faith. 
            19                  Again, sometimes -- not in your 
            20       particular hypothetical -- you might know that 
            21       things would be ignored.
            22   Q.  What do you mean?
            23                  MR. GARBUS:  It seems to me to speak 
            24       for itself.



             1                  MR. HART:  Well, I'm sure it does, but 
             2       I want to know what the witness means by that. 
             3   A.  I know an example, and I would rather not give 
             4       names, of someone who discovered that because of 
             5       the security system of a new computer, you could 
             6       remotely turn on the microphone in that computer 
             7       and just hear everything that was happening in 
             8       the person's office.  And the CEO of this company 
             9       was someone who usually didn't listen to any 
            10       kinds of suggestions, and I know the person who 
            11       got the CEO's attention by handing him a tape 
            12       recording of everything he said in the office for 
            13       the past three days.  That's an example of, I 
            14       think, really informing someone that you've had 
            15       an effect.
            16   Q.  That was a more dramatic way of informing the CEO 
            17       in your example, right?
            18   A.  Right.  This wasn't a computer on the market, it 
            19       was just something they were planning. 
            20   Q.  Now, am I correct in understanding that it's up 
            21       to you whether or not you will be available for 
            22       the trial in this case?
            23   A.  It's more up to Hewlett Packard.  It depends on 
            24       what kinds of meetings get scheduled, then 



             1       depending on what types of meetings, it's more or 
             2       less difficult for me to be able to appear. 
             3                  MR. HART:  I don't think I have 
             4       anything further.  Thank you again. 
             5                  MR. GARBUS:  Thank you. 
             6                  (Whereupon at 1:38 p.m., the 
             7       deposition was adjourned.)



             1                  C E R T I F I C A T E
             2       I, HAROLD ABELSON, do hereby certify that I have 
                     read the foregoing transcript of my testimony 
             3       given on July 5, 2000, and I further certify that 
                     said transcript is a true and accurate record of 
             4       said testimony (with the exception of the 
                     corrections listed below):
                     Page      Line       Correction
            16       Dated at                         , this        
                     day of                 , 2000.
            18                              HAROLD ABELSON



             1                  C E R T I F I C A T E
             4       SUFFOLK, SS.
             5               I, Loretta Hennessey, Registered Merit 
             6       Reporter and Notary Public in and for the 
             7       Commonwealth of Massachusetts, do hereby certify:
             8               That HAROLD ABELSON, the witness whose 
             9       testimony is hereinbefore set forth, was duly 
            10       sworn by me and that such testimony is a true and 
            11       accurate record of my stenotype notes taken in 
            12       the foregoing matter, to the best of my 
            13       knowledge, skill and ability.
            14               IN WITNESS WHEREOF, I have hereunto set 
            15       my hand and Notarial Seal this 6th day of July, 
            16       2000.
            20                              Loretta Hennessey, RMR
            21                              Notary Public
            23       My Commission Expires: 6/10/05