The trial against 2600 Magazine commenced at 9am today (Monday) in the
federal court house at 500 Pearl in NYC. Throughout the day,
approximately 40 protestors stood behind a police blockade with anti-MPAA
and anti-DMCA signs, chanting some great slogans. The court room was
packed all day.
Unfortunately, I did not arrive until approximately 10:30am and thus
missed hearing discussion of defendant's motion to recuse/disqualify Judge
Kaplan, as well as opening arguments. Text of the motion/supporting memo
have not shown up on Cryptome yet, but can be found at
http://www.eff.org/IP/Video/MPAA_DVD_cases/
<disclaimer>If I'd been allowed to make use of my laptop during the trial,
I would not have had to resort to frantically scrawling 12 pages of
shorthand questions and answers on 1cm-lined paper. I also could have
simply posted my notes in their original form without need for a
reinterpretation. I hope what follows makes sense. If not, I'd be happy
to elaborate/clarify as best I can.</disclaimer>
The bulk of the day focused on the testimony of Michael Shamos, a faculty
member in the CMU computer science department. Shamos is also an IP
attorney. In his June 30, 2000 declaration, Shamos claimed he was hired
by the MPAA attorneys to "conduct certain experiments to investigate the
feasibility of obtaining DVD decryption software (DeCSS), using that
software to decrypt commercial DVDs, converting the decrypted videos to
DivX format and transferring those DivX files between computers on local
area networks." Note the words "experiment," "investigate," and
"feasibility". They're important.
I strongly recommend that you read Shamos's declaration. (It's on
Cryptome.) IMO, this man is a disgrace to academic computing. This
should be readily apparent from the following, and from examination of his
declaration. (Also, a lot of the stuff below will not make sense unless
you've read the declaration.)
In case you don't make it to the end of this document, Frank Stevenson
will be testifying at 9am on Tuesday. (I'm greatly looking forward to
just how badly the "is that megabits or megabytes?" MPAA lawyers will
cross-examine a competent cryptographer.)
----------
Around 10:30, Shamos was asked by Simms (the MPAA lawyer) to compare
quality differences between a "legit" DVD, and a DivX created in part via
DeCSS. The MPAA had set up a monitor for this demonstration, but after
protestations from Garbus (that a television and a laptop screen were not
the same thing), Judge Kaplan suggested everyone interested just watch the
demonstration right on Shamos's laptop. Some 15 people subsequently
hovered around the laptop -- MPAA lawyers, defense lawyers, a little girl,
the court reporter, John Gilmore, and a federal judge in robes...all
bunched in the middle of the court room to view the opening credits of
Sleepless in Seattle. It was a memorable sight.
Shamos proceeded to point out objects in the foreground of the movie,
noting their clarity and color and such. At the direction of Simms,
Shamos's assistant (Eric Burns) then played the DivX file version. Shamos
pointed out the same objects, this time explaining that he was looking for
"artifacts"... irregularities on surfaces which were uniform in the
original DVD. He explained that they were extremely difficult to spot,
and then embarked down a path which I hope will ultimately disqualify him
as an "expert" in this case.
Shamos: "We've conducted numerous lay experiments where we've asked
people to come in here and tell if it was a DivX or a DVD..."
Judge Kaplan: "That's obviously hearsay."
Shamos was then asked a round of questions by Simms pertaining to whether
he attempted to obtain/trade DivX files with others. Garbus repeatedly
objected, as much of the "experiment" was ultimately done by Shamos's
assistant. Garbus also frequently objected because Simms failed to "lay a
foundation" establishing how Shamos was qualified to answer the questions.
Shamos proceeded to recount what was covered in his declaration...sort
of. He explained how he and his assistant logged an IRC session in which
they traded a Sleepless in Seattle DivX file for a Matrix DivX file.
When asked by Simms how long it took to transfer the Matrix DivX file,
Shamos could not recall. He estimated approximately 6 hours, but stated
that he did not know either the start or conclude time of the transfer.
Shamos was then asked by Simms about the "technical parameters" of the
transfer. Shamos estimated the effective bitrate of the transfer at 250
kbps and stated the internet connection was via a 100 megabit switch
network.
[10 minute break]
After the break, Shamos was asked by Simms to repeat his DVD vs. DivX
quality demonstration with copies of the Matrix. Once again, some 15
people crowded around the laptop and watched clips of both a legit Matrix
DVD, and a Matrix DivX file. Garbus objected to the demonstration,
stating that unless one had seen the Matrix 35 times, it would be
impossible to tell the difference between the two without a side-by-side
comparison.
Judge Kaplan: "This is not Gone With the Wind."
At Simms' prompting, Shamos then discussed how he was able to transfer the
Matrix across the LAN in 20 minutes (10bT) and 3 minutes (100bT). After
that, he answered a number of questions pertaining to his "qualifications"
for discussing the availability of bandwidth at universities and in the
country in general.
Shamos: "There is a massive need for additional bandwidth."
Next, he was asked by Simms what would happen if a large number of
students attempted to trade DivXs simultaneously. Shamos stated that the
network would become slower and slower. Simms then asked how network
traffic fared at the time of the experiment. Shamos stated that he
effectively had about 1/3 of the theoretical maximum bandwidth.
[lunch until 2:15]
After lunch, Shamos testified that since January of 2000, approximately
650 movies have been converted into DivX format, comprising approximately
10% of the available DVDs. He also stated that obtaining DivX files was
easy over IRC, and more difficult if one contacted people through iSONEWS.
His testimony for the MPAA concluded with comments that DeCSS played a
major role in the making of DivX files because it was recommended as the
"tool of choice" on multiple web sites, "easy to obtain," and "easy to
use." When asked if there was an easier way to descramble a DVD, Shamos
stated that he was not aware of any but admitted not having "deep
familiarity" with content of the question. (Ding ding ding.)
Both sets of lawyers got into an argument at the close of Shamos's direct
examination, which I couldn't note fast enough...
Judge Kaplan: "I'm not sure I understand what you're all saying to me."
---
Before the defense began cross-examination of Michael Shamos, Hernstadt
(one of the defense attorneys) informed Judge Kaplan that Frank Stevenson
was in town and would be returning to Oslo Tuesday evening. Hernstadt
asked whether Stevenson could be permitted to testify. Kaplan said that
he would hear discussion on that at 4:30, after Shamos was excused.
---
I was told by several people that Garbus's performance earlier in the
morning (re: the recusal motion and opening statements) was sub par. I
didn't witness it, so I don't know. I can say that IMHO he did a
wonderful job in the afternoon.
Garbus began by eliciting the conditions under which Shamos undertook his
"project" for the MPAA attorneys. Shamos previously made $25k as an
expert witness in iCrave. He estimated that he would make $30k from the
MPAA for his current work; his rate with the MPAA is $400 an hour.
[Garbus's cross-examination skipped around to different topics...I'm not
going to alter their order nor will I attempt to reword the following into
enjoyable reading...I'm trying to keep this as true to a prose rewrite of
my shorthand transcript as possible.]
Shamos was sent ~300 pages of affidavidts/depositions by the MPAA
attorneys. When questioned, Shamos remembered reading MIT prof. Harold
Abelson's declaration, as well as that of CMU prof. Dave Touretsky.
Shamos recalled that Touretsky hosted a web page which contained various
incarnations of DeCSS, including C source, an image of DeCSS on a tshirt,
the DeCSS algorithm, an anotated version of the algorithm,
etc. (note: he knew this BEFORE his 'experiment'.)
Garbus then asked if Shamos was aware of how many people accessed
Touretsky's page. Shamos was not aware and suggested that Garbus could
examine the CMU logs.
Garbus then asked Shamos to explain how he decided to become involved in
the case. Shamos said that after receiving the court documents, and after
the preliminary injunction was issued against 2600, he decided to
participate.
Next, Garbus asked Shamos the following questions (not sure if these are
exact quotes...but you get the general idea):
Garbus: Do you know of one person who has copied a DVD as a result of any
download from 2600?
Shamos: No.
Garbus: Has anyone at MPAA told you of anyone having copied a DVD [from
2600]?
Shamos: No.
[etc. Shamos admitted that no one told him the DivX of the Matrix for
which his assistant traded a DivX of Sleepless in Seattle was made with
help of DeCSS.]
Garbus moved on to questions about the resources of the MPAA and whether
the MPAA had conducted any studies on DeCSS. [don't know; no] Then, to
when CSS was first cracked. [1997] Then, whether he'd seen documents
given to Schumann [no], then back again to how much money the MPAA spent
trying to determine who copied DVDs. [no way of knowing]
Shamos then explained that DVDs would replace VHS because they allowed
users to "skip around" and "are digital."
Shamos was asked to do his "study" by William Hart. It was specifically
requested for this lawsuit, and Shamos was selected because of his
expertise in the area. Shamos admitted to being given a precise set of
instructions on how to carry out his study. [I missed some things
here...they were talking quickly...something about being asked by Hart to
remove references to the 1st Amendment and to the NYT's linking to DeCSS
from his early declaration drafts to avoid "clouding the case with 1st
Amendment issues."]
Garbus next asked if Shamos had read any issues of the Hacker Quarterly.
[no] The line of questioning wandered off to how, when there were
thousands of sites (according to hotbot) which contained references to
DeCSS, Shamos opted to grab DeCSS specically from 2600's site. Answer:
***Shamos was told to do so by the MPAA lawyers.***
[questions skipped around again: the "study" required several
steps; Shamos had to locate #divx as one of them; Shamos has used
computers since 1962; Shamos spends 4 hours a day on computers; Shamos
teaches corporate classes on the side; Shamos makes $700k so
doing; courses concern ecommerce]
Next, Garbus asked about the qualifications of Shamos's assistant on the
project, Eric Burns. Eric is a CS undergrad at CMU.
Garbus inquired about whether or not school was in session during the
period when the study was conducted. 4/5ths of the school was not in
session. (CMU adjourns sometime in May.) Shamos admitted that network
usage during the periods when school is not in session is not the same as
when it is. Shamos admitted that he did not include this in his report
for the MPAA. The MPAA did not ask Shamos about normal usage. Shamos did
not know what the results would have been had the "study" been conducted
during the school year. Shamos admitted that some of the "study" was
connected on the network after midnight, that he knew there was a
difference in network speed at that time, and that he also did not include
this information in his report.
[break]
Shamos explained that he and Eric purchased a computer for the test around
3pm of some day which he could not recall. They took it back to the CMU
CS offices, set it up, and connected it to the network. [in case you
haven't read the Shamos declaration, they also bought a copy of Sleepless
in Seattle on DVD]
Garbus listed several alternatives to DeCSS, asked if Shamos had heard of
them, or if Shamos had asked the MPAA if there were faster copy techniques
than any including DeCSS. [answer to all questions was no.] Interspersed
in these questions were rehashed questions of Eric's qualifications as an
assistant.
Next, Garbus asked about LAN speeds at CMU 2 years ago [same as now], 4
years ago [don't know], and the general trend [don't know].
Garbus: To your knowledge, isn't it true that LAN speeds haven't changed?
Shamos: [don't know]
Shamos was not surprised to hear that Northern Lights found 700k hits for
DeCSS. Shamos did not know how many sites in the US contained DeCSS in
object or source code, nor did he know outside the US. Google had a large
number of DeCSS hits. Shamos did not search Yahoo or AOL for them.
When asked the speed of the Internet -> CMU connection, Shamos called it
"very fast." When asked what very fast meant, Shamos said, "We try to
have the best stuff," and explained that CMU was one of the leading
schools and therefore had the best connections. (Not bad for an
"expert," eh?)
After downloading DeCSS via 2600, it took about 45 minutes to get the
Sleepless in Seattle DVD DeCSS'd. Next, they attempted to make it into a
DivX. It was about 10pm on the first day of the "experiment" by this
time. At 10pm, Eric and Shamos began the process of learning how to
compress into DivX. Shamos had not done this before.
When Garbus asked whether Eric had made a DivX before, he pointed out that
as there was a possibility Eric would be called to testify, Garbus
strongly hoped Eric was not currently present in the room.
A sheepish-looking Eric Burns immediately stood up from the back of the
court room and exited.
Garbus asked again whether Eric had made a DivX before and Shamos said
that he didn't know, and that the goal was not to make implicit use of
Eric's technical expertise.
Shamos then explained that the whole process was done in several parts,
the hardest of which was sync'ing the audio and video tracks. Shamos had
no previous experience in sync'ing. He was not sure if Eric had. Shamos
said the sync'ing process took them approximately 10 hours of trial and
error.
Garbus: Would you agree it's tough?
Shamos: It's not fun.
Garbus: It's difficult?
Shamos: It's a lot easier than making the movie in the first place.
The questioning went back to the sequence of events. Shamos could not
recall what happened at 10pm on the first day of the "study." He admitted
to not having any notes on when Eric left on Day 1, nor to having logs or
timesheets of specific hours spent on the project or what was
accomplished. The MPAA lawyers did not suggest such timesheets.
Garbus skipped ahead to the IRC session where the Sleepless in Seattle
DivX was traded for the Matrix. He asked if the logs provided of that
session were complete. [they were not.] There were two logs made -- one
of the general #divx chat, and one made of a private conversation
initiated in #divx.
Garbus revealed that ***some of the log was typed in.*** This was not in
Shamos's declaration, nor was Hart informed of it.
Garbus's questioning returned to the point in time in which the DivX was
sync'd. Shamos did not know what time the sync'ing was finished. He said
Eric told him as soon as it was complete; they double checked that the
sync was proper; then they started looking for a trade.
Shamos stated that around 11pm, Shamos and Eric were in Eric's office.
Eric began looking for a trader, and when he found one and began the
transfer, Shamos went back to his own office.
Shamos did not know the speed of the connection, but surmised it was
1/3. Garbus then asked how long the IRC chat took. Shamos did not
know. Judge Kaplan pointed out that the log covered 23:15 to 23:45 and
asked whether that was valid.
Shamos: The times in the log have not been altered.
[Garbus then pointed out an inconsistency between P.245 and the time when
Shamos claims to have left. I didn't write down what this was...]
Garbus then asked whether Shamos kept records when experimenting at
CMU. Shamos stated that he didn't do experiments.
Garbus asked when Shamos left on the next day, and asked whether the MPAA
lawyers were billed. [4am, i don't know] Shamos did not have
timesheets. Shamos has not billed the MPAA lawyers yet. His desktop
computer has a small log of hours spent per day on the project, but it
lacks specificity. Eric makes $100 an hour from the MPAA lawyers.
Shamos stated that Eric attempted to DivX Fight Club, but was not
successful. Shamos did not know why the attempt failed, nor was he aware
that DeCSS did not work on Disney movies. Eric was not asked to copy
Fight Club, nor were there any memos or notes about the matter. Shamos
could not remember telling the MPAA lawyers that DeCSS failed on Fight
Club, but stated that it sounded logical that he would have done so.
Garbus concluded cross-examination with a handful of random questions, the
last of which was whether or not CMU approved of his experiment. Shamos
very defensively said it was acceptable use of the university computer
systems. Garbus said he would ask Touretsky's opinion on the matter
later.
---Redirect of Shamos (MPAA lawyers try to cover their wounds...)
For the portion of the experiment which involved use of DeCSS, the laptop
was not connected to the LAN. LAN traffic does not affect processing
time.
---End redirect.
Garbus then stated his desire to move to exclude Shamos as an expert.
Kaplan postponed that discussion until later, and then heard some
discussion regarding the deposition/testimony of Frank Stevenson. I'm not
going to bother with my transcription of that. Frank's testifying
tomorrow morning and I've only got 1.5 hours to sleep. =)
