IN THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF PENNSYLVANIA - - - AMERICAN CIVIL LIBERTIES : CIVIL ACTION NO. 96-963-M UNION, et al : Plaintiffs : : v. : Philadelphia, Pennsylvania : March 21, 1996 JANET RENO, in her official : capacity as ATTORNEY GENERAL : OF THE UNITED STATES, : Defendant : . . . . . . . . . . . . . . . . HEARING BEFORE: THE HONORABLE DOLORES K. SLOVITER, CHIEF JUDGE, UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT THE HONORABLE RONALD L. BUCKWALTER THE HONORABLE STEWART DALZELL UNITED STATES DISTRICT JUDGES - - - APPEARANCES: For the Plaintiffs: CHRISTOPHER A. HANSEN, ESQUIRE MARJORIE HEINS, ESQUIRE ANN BEESON, ESQUIRE American Civil Liberties Union 132 West 43rd Street New York, NY 10036 -and- STEFAN PRESSER, ESQUIRE American Civil Liberties Union 123 S. 9th Street, Suite 701 Philadelphia, PA 19107 -and- For the ALA BRUCE J. ENNIS, JR., ESQUIRE Plaintiffs: ANN M. KAPPLER, ESQUIRE JOHN B. MORRIS, JR., ESQUIRE Jenner and Block 601 13th Street, N.W. Washington, DC 20005 - - - APPEARANCES: (Continued) For the Defendant: ANTHONY J. COPPOLINO, ESQUIRE PATRICIA RUSSOTTO, ESQUIRE JASON R. BARON, ESQUIRE THEODORE C. HIRT Department of Justice 901 E. Street, N.W. Washington, DC 20530 -and- MARK KMETZ, ESQUIRE U.S. Attorney's Office 615 Chestnut Street, Suite 1250 Philadelphia, PA 19106 - - - Also Present: MICHAEL KUNZ Clerk of the Court for the Eastern District of Pennsylvania - - - Deputy Clerks: Thomas Clewley Matthew J. Higgins Audio Operator: Andrea L. Mack Transcribed by: Geraldine C. Laws Grace Williams Tracey Williams Laws Transcription Service (Proceedings recorded by electronic sound recording; transcript provided by computer-aided transcription service.) (Whereupon the following occurred in open court at 9:32 o'clock a.m.:) CLERK OF COURT KUNZ: Oyez, oyez, oyez, all manner of persons having any matter to present before the Honorable Dolores K. Sloviter, Chief Judge of the United States Court of Appeals for the Third Circuit and the Honorable Ronald L. Buckwalter and the Honorable Stewart Dalzell, Judges of the United States District Court for the Eastern District of Pennsylvania may at present appear and they shall be heard. God save the United States and this Honorable Court. Court is now in session, please be seated. COUNSEL: Good morning, your Honor. JUDGE SLOVITER: Good morning. This is the hearing before the statutory Three-Judge Court on the request for a preliminary injunction in the consolidated action of American Civil Liberties Union and its co-plaintiffs versus Reno, No. 96-963, and American Library Association and its co-plaintiffs versus Department of Justice, No. 96-1458. Judge Buckwalter and I want to thank Judge Dalzell for his case management of all preliminary matters and all three Judges want to thank the parties for their concerted efforts to expedite the proceedings by entering into stipulations and we want to thank the plaintiffs for presenting their testimony in chief by affidavits which are available for public and press view. We -- now, there have been various requests for photographing the proceedings and I thought it might be, we thought it might be appropriate to make a statement at the inception. In September 1994 the Judicial Conference of the United States voted not to permit the taking of photographs and radio and television coverage of proceedings in the United States District Courts, whether those proceedings are civil or criminal. When the Judicial Conference voted last week to permit each Court of Appeals to decide whether and under what circumstances to permit the taking of photographs and radio and television arguments of appellate arguments it reiterated its prior policy with respect to the District Courts. This proceeding is a three-judge District Court proceeding. I am a member of the Judicial Conference of the United States and Chief Judge Cahn of the Eastern District of Pennsylvania is a member of the Judicial Conference. Whatever the views of individual members of the Eastern District on the issue of cameras in the Federal courtroom or the collective views on that issue of that District Court which was one of the pilot courts during the project and whatever the personal views of the Judges on this panel on that issue, we will abide by the vote of the Judicial Conference of the United States, whether binding or hortatory. We are privileged to be part of the great institution that is the Federal Judiciary of the United States. Okay, we will proceed with cross-examination of the plaintiff's witness by the Government and we ask the Government at the inception is there any -- we understand that the first witness is -- and we'll ask the Government, whom are you calling? The Government. (Laughter.) JUDGE SLOVITER: The Court of Appeals, go ahead. MR. COPPOLINO: I'm sorry, your Honor, your question? JUDGE SLOVITER: Your first witness? MR. COPPOLINO: Would be Scott Bradner, I believe, your Honor. JUDGE SLOVITER: Okay. And do you accept Mr. Bradner as an expert? MR. COPPOLINO: Yes, we do, your Honor. JUDGE SLOVITER: Okay, thank you. Then proceed. JUDGE DALZELL: Oh, yes, Mr. Coppolino, just one housekeeping, important housekeeping matter. We were most grateful to the parties for the extensive stipulations that were submitted that will probably save us a week of testimony but there was -- and I don't criticize you for this -- there was a qualification at the beginning of the stipulating having to do with the fact that discovery of the plaintiff's case was not concluded and therefore it says that no party waives its right to submit information inconsistent with its terms should it learn thereof through discovery, disclosure or other investigation. Now, I take it that by April 1st, that is to say when the plaintiffs complete their case, that the Government will be in a position to tell us yea or nay whether it indeed accepts every paragraph of those stipulations. MR. COPPOLINO: Your Honor, I believe I can represent that the purpose of the stipulation on the part of both parties was to indicate that it was applicable to the entire PI hearing. And I believe that the specific purpose being addressed here is that as the case proceeds beyond the PI stage of the merits and discovery is taken at that time which indicates that some of the statements in here are incorrect or inaccurate upon further testing that at that point these particular stipulations would not apply. That's my understanding. JUDGE DALZELL: All right. So therefore you can represent to us now that for purposes of this preliminary injunction hearing that we may take the stipulation without qualification. MR. COPPOLINO: Yes. JUDGE DALZELL: Okay, fine, that's very helpful. Thank you. Okay, do you want to -- it's Mr. Bradner, is it? MR. MORRIS: Yes, your Honor, my name is John Morris, co- counsel for the ALA plaintiffs and plaintiffs call as their first witness Scott O. Bradner. MR. KMETZ: Your Honor, Mr. Jason Baron will be handling the cross-examination. JUDGE SLOVITER: And it's our understanding that there will be only one lawyer per witness. MR. KMETZ: That would be our understanding as well. MR. MORRIS: That's certainly our understanding. If the Court would indulge at the conclusion of any redirect we might have, I will confer just momentarily among ourselves to make sure that we're all on the same page. JUDGE SLOVITER: We don't mind your conferring as long as you don't mind our conferring -- (Laughter.) JUDGE SLOVITER: -- because a three-judge court is something new for all of us, three-judge District Court. THE COURT CLERK: Sir, will you state and spell your full name for the record? THE WITNESS: Scott Bradner, S-c-o-t-t B-r-a-d-n-e-r. THE COURT CLERK: Will you place your left hand on the Bible and raise your right hand? SCOTT BRADNER, Plaintiffs' Witness, Affirmed. MR. MORRIS: And at this point the plaintiffs would move into admission the evidence of the previously filed declaration of Mr. Bradner as sworn to on the 19th of this month as his trial testimony. And Mr. Bradner is available for examination by the Government and certainly any questions the Court may have, I'm sure he'd be happy to respond to. JUDGE SLOVITER: Thank you. Is there any objection to-- MR. BARON: No objection, your Honor. JUDGE SLOVITER: -- accepting that as evidence, fine. Proceed. MR. BARON: Good morning, your Honors. JUDGE SLOVITER: Good morning. JUDGE DALZELL: Good morning. JUDGE BUCKWALTER: Good morning. CROSS-EXAMINATION BY MR. BARON: Q Good morning, Mr. Bradner. You state in your decla-- JUDGE SLOVITER: Excuse me. In the Court of Appeals we always identify ourselves, we ask the counsel to identify themselves for the record. Maybe that would be a good idea. MR. BARON: My apologies, your Honor. JUDGE SLOVITER: That's all right. MR. BARON: I'm Jason R. Baron, B-a-r-o-n, counsel to the U.S. Department of Justice. JUDGE SLOVITER: Thank you. MR. BARON: Thank you, your Honor. BY MR. BARON: Q Mr. Bradner, you state in your declaration that you are co- area director of something called the IETF. Could you, in a nutshell, tell us what the IETF is and what does it do? A The Internet Engineering Task Force is a self-organizing group which developed out of some U.S. Federal Government networking initiatives many years ago and it is the group which now is primarily responsible for developing standards for use in the Internet protocol which is the basis upon which the Internet runs. Q The IETF has been in existence for about ten years, correct? A That is correct. I'd be clear that it predates my involvement so I'm taking that from what others have said. Q Okay. Would it be fair to say that the IETF defines standards for the Internet Protocol suite? A That is -- yes, it would be fair to say that. Q Could you explain for the Court what is the Internet Protocol, otherwise known as IP? A The Internet itself consists of many networks connected together by other networks. The Internet Protocol is that part of the protocol suite, that part of the language which is used on the network which allows a piece of information called a packet on one network to find its way to identify a separate network and find its way to that separate network. So the IP is the Internet Protocol that allows movement of data between networks. Q Different protocols make up IP suite, correct? A That is correct. Q Can you name a few for the Court? A Well, the underlying protocol is the Internet Protocol or IP. Riding on top of that are protocols such as TCP, the Transmission Control Protocol, UDP, the Unreliable Datagram Protocol, ICMP, the Internet Control Message Protocol. Riding on top of TCP are protocols such as Telnet on the World Wide Web, HTTP Protocols. It's a layer cake of various different concoctions. Q Well, we'll get into some of those in a few minutes. Who comprises the IETF? A As I said, the IETF is a self-organized group, we have meetings three times a year. The membership is those who attend the meetings and those who are on the mailing lists. There are some 80 or so working groups, each of the working groups maintains a mailing list and anybody who joins any of those mailing lists is de facto part of the IETF. There is an organizational structure within the IETF which divides the working groups up into areas and then the area, each area is managed by one or more, one or two area directors. Q We're going to get into that as well. People doing the, quote, "standards work," unquote, on the IETF, are they normally paid by corporations and businesses? A Or they are paid by corporations or businesses or universities or their private consultants. Q Okay. Could you explain to the Court what an RFC is? A RFC came from the original process of asking for comments, it stood for request for comments, asking for comments on thoughts on how to do some proposal. It is progressed past that point now and RFC stands for RFC. It no longer is a vehicle for comments. There is a new vehicle for the comments which are called Internet drafts and they pre--precede RFC's, but RFC's are the basic standardization document series for the IETF. Q RFC's exist that define a standards process for the Internet, correct? A There are a series of RFC's which have progressively defined the standards process. Q And some of the RFC's establishing a standards process for the Internet are well established, correct? A There is -- the original standards process was defined in RFC- 1310, that has been superseded by RFC-1602 which has been in effect for a few years, I don't remember exactly what. And within one of the working groups within the IETF is called the Poise Working Group -- and don't ask me what that stands for cause I don't know -- and that is in the process of refining a third revision of the standards process. It's now known as 1602 BIS because it has not gotten its own RFC number yet. That should happen within a few weeks. Q You're editing at least one RFC at the present time having to do with Internet standards, correct? A I am editing two of them; co-editing one and editing another one. Q Now, you mentioned an area within IETF, what is an area within the IETF? A An area is a grouping of working groups, normally trying to make the -- it's normally tried to be done in a way which is logical so that the working groups which are working on security- related matters are grouped in the security area. Working groups that are working on network management related efforts are in the network management area. Q Would it be fair to say that the area of operational requirements that you are co-area director of has to do with developing standards for the next generation of software for the Internet? A No, it would not be. The operational requirements area is a little bit of a confusion point on the IETF in that one of the things that we feel we must have is some kind of feedback from the operation of a protocol to the protocol developers and the operational requirements area does two things: one, tries to make sure that when standards are developed they can be done so, the resultant standards can actually be operated in the real world rather than just in the theoretical world. And then, second of all, if indeed when these standards are deployed that there is any lessons to be learned which should go back to the standards developers that they are fed back. I think what you may be referring to is the ad hoc or the temporary IP Next Generation or IPNG area which is, was working on and is currently working on extending for the new generation of the IP protocol itself. Q Well, you have also been the co-area director of the IP Next Generation area, correct? A That's correct. Q It sounds a little bit like Star Trek; what does that area consist of? A It was purposely it sounds a little like Star Trek, actually. (Laughter.) A It consists of the -- Allison Menkin and I were asked by the -- by Phil Gross, the chair of the IETF, to put together a temporary area to -- to group together all of the activities involved, all of the proposals for a successor protocol to IP to deal with scaling issues and the like and to resolve the question of what should be "The Proposal" out of the IETF-4 and IP Next Generation. There were a number of proposals on the books when we were assigned the task of forming this temporary area, we have made a recommendation on what the next generation should be, that recommendation has been accepted and the area right now is closing down because it's very close to have finished publishing the initial set of RFC's, the initial set of standards for IP Next Generation. Q Would it be fair to say, to summarize what you've just said, that the IP Next Generation group is working on a new generation of the IP Protocol itself? A That is correct. Q Does it have -- does the IP Next Generation group have recommendations regarding a specific architecture of the packet traffic on the Internet, including the format of the packet? A It has a recommendation on the format of the packet itself that's actually the basic recommendation is the format of the packet traffic itself. You used the word "architecture" in your question and that's potentially confusing because architecture could mean the way the networks are put together, it could mean the concept of how the packets are flowed through the network, it could mean a number of different things, so I would prefer to say that we've defined the packet format itself and we have looked at architecture in various areas but not come to a specific recommendation on architecture. JUDGE DALZELL: Sir, when you -- excuse me. When you use the word "architecture" and it's in all -- a number of the declarations, there's no -- that's not a term of art that means one thing in this area? THE WITNESS: It means -- it means one thing for each of the areas that it's in. And it's as a security architecture which ties together a unified view of how one should do security. There is a routing architecture which ties together a unified view of how one should do routing which is the keeping track of where networks are. So there are a number of architectures depending on what particular topic we're on. There isn't an overall architecture because at the moment it's too complex a network with too many functions going on. You have to look at the individual functions and do an architecture on those. We have done some work in that area, there is an architecture for security, an IP Next Generation Security which is now the general IP Security architecture and we have looked at architecture in other areas, but it's difficult to do to unify all of the thoughts together. One of our recommendations in the -- in our recommendation for IP Next Generation was to appoint an individual to be an architect for IP Next Generation. Unfortunately, there aren't many people who could do that task and fewer of them with enough time to do it. So we in the -- as Allison and I have acted as architects to make sure that there's a consistent view of what IPNG, IP Next Generation looks like across the various activities creating protocols for it, the TCP Next Generation, the ICNP Next Generation, the routing, security, all of these different working groups working on different aspects of it, we're trying to keep their view of what IP Next Generation looks like as consistent. So we in that context are acting as architects, but it's architects is one of those words which depends on the beholder. JUDGE DALZELL: Depends on? THE WITNESS: On the beholder. JUDGE DALZELL: Okay. But you consider yourself one? THE WITNESS: In a real sense, no, I do not. Architects tend to be more visionary than I tend to be in this environment, intend to be more on intuitive feeling of how the incredibly complicated world of the networking fits together and what the implications are of making a change some place. I think that I can understand architecture, but I would not go so far as to say that I am an architect in the context of, for example, Dave Clark of MIT, Dr. Dave Clark, who is -- who was the original IP architect and the one who--and the person that unfortunately didn't have enough time to be the IP Next Generation architect. JUDGE DALZELL: But, for example, in the stipulation and we hear a lot about the packet switching, for example. Now, would that be like, to continue the metaphor, a brick that is commonly used in all forms of architecture? THE WITNESS: It's -- maybe it's more of a fact of life of the forms of architecture, but -- JUDGE DALZELL: Because that's unchanging, the packet switching concept, that's not going to change? THE WITNESS: One of the areas that we specifically addressed in working on IP Next Generation was what are the paradigms which we want to follow in IP Next Generation and one of them was we wanted to preserve what is called the Datagram mode which is the packet mode. The alternative to that is circuit switching like a telephone where you do a call setup, you do initialization and all of the traffic flows down a particular path. The original IP that designed when it was originally designed was designed to deal with adverse events. The colloquial story is it was designed to deal with atomic war which is an adverse event. And -- JUDGE DALZELL: I think we can agree on that. (Laughter.) THE WITNESS: And so the ability for IP to survive that kind of environment, the kind of hostile environment we felt was very important to maintain, Datagram mode means that the individual units of the data that move over the network which are packets have full identify-- source identifiers and destination identifiers in each packet and are separately routed, separately handled by the computers which comprise the network, therefore being resilient to individual failure. JUDGE DALZELL: And that is -- but that is not subject to change right now, that's not in the NG, the IPNG? THE WITNESS: We specifically, specifically chose to require the support for Datagram mode in IPNG. JUDGE DALZELL: Thank you. Sorry to interrupt. BY MR. BARON: Q Mr. Bradner, are all IETF documents public? A It is a -- it is a matter of pride and honor in the IETF that all documents are public documents available for free over the net. We used the paradigm to develop the paradigm. Q And that includes all RFC drafts or proposals for standards, right? A That is correct, they are called Internet Drafts and they are publicly available. Q And they're put up on web sites and are available to the world at large, correct? A That is correct. Q Can we pause here and define what a URL is for the Court? A URL is a term which means Uniform Resource Locator, a pointer. It's the best -- the best way to identi-- to consider it is it's sort of like a combination of all of the things you might have in a phone directory listing, somebody's name and address, and it is where something is on the Net, not relative to you but in an absolute sense. You don't go three buildings over to the left and two stores down, it is here is the absolute location of something ir-- independent of where you happen to be sitting in the network. Q For the IETF itself, am I correct that the URL is something known as HTTP://WWW.EFF.ORG? A That is incorrect. Q Oh. A Now, I'm sure that -- Q Please correct me. A -- you -- you meant that to be incorrect. Q Oh, I see, yes, no, I'm sorry, I -- I did not have the right URL. Why don't you give the right URL? A HTTT -- HTTP://WWW.IETF.ORD. MR. BARON: Yes, your Honors, I think I've been reading the EFF site on the Web too much. BY MR. BARON: Q Okay. Could you explain what these domains in that URL represent for the IETF URL? A The -- the part which is relative to the IETF is the WWW.IETF.ORD. The part which precedes that, HTTP, is the protocol, the function in which one should retrieve, should access this site. Different options there are, for example, FTP for File Transfer Protocol or Gofer, are different kinds of concepts, different kinds of application programs to use to access this site. In this particular case where you can access that particular site with FTP with anonymous FTP or with the Web, WWW, the URL you specified is one which is using the Web to access this site. Q What's the difference between the current IP Version Four and the Next Generation Version Six of the protocol? A How much detail would you like that answer in? Q Oh, just -- just sort of a summary for the Court. A The reason to undertake the effort, and it was some significant effort to develop a new generation for IP revolved around three basic issues: The first issue was that the IP address itself which is the field, I mentioned in the packets themselves there is this source identifier and a destination identifier, a source address and a destination address. In IP Version Four which is the current version, those fields are 32 bits long, each of which could in theory identify four billion individual posts or computers on the Network, but because of address assignment inefficiency we're beginning to run out of those and we're beginning to run out at a rate which caused a great deal of consternation, particularly in the press back in the early 90's, '92 and '93, that investment in IP was probably not a good idea because we were running out of addresses. It's like going to the phone company and saying I'd like a phone and they say they don't have a number. So the first thing was to try and fix the problem of running out of addresses. The second thing was to try and fix the problem of that there was too much routing information, this is the information within the computers that tie the Internet together, they're called routers, they're special purpose computers. And in each one of those computers in the backbone, in the more central locations within the network must keep track of where every network in the world is, every -- the network which is the one which connects this computer here has to be kept track of by those computers and the routers in the backbone. The size of the routing table, the size of that information was growing faster than memory technology, doubling every nine to ten months and memory technology was doubling every 11 to 24 months and in the long run those two lines will never intersect. And so we had to do something. JUDGE DALZELL: Excuse me, it's doubling every nine to twelve months? THE WITNESS: In nine -- JUDGE DALZELL: It's doubling every nine to -- THE WITNESS: The size of the Internet, yes. JUDGE DALZELL: Nine to twelve months? THE WITNESS: Yes. It's been tending towards the nine month area of doubling. And then the third area was want to be able to deal with improving some aspects of the current Internet, security aspects, real time or flow control or quality of service metrics and things like that, so those were the three areas which we were focusing on. In the first area the IP Version Four address, as I said, is 32 bits long. The IP Version Six which is what IP Next Generation is, is 128 bits long. Now, that's four times the number of bits, but that's actually four billion times four billion times four billion times the number of hosts. That turns out to be a very large number, yet somebody estimated that even under the absolute worst efficiency, the least efficient method of allocating them it still works out to 1500 computers per square meter of the earth's surface, including the oceans. (Laughter.) THE WITNESS: We think that we have -- we think that we have aimed for the future in the expandability of this. (Laughter.) BY MR. BARON: Q Sounds like a pervasive number of bits. A Okay. The second, the second area was dealing with the routing table space and we've made the addresses aggregatable so that instead of having to articulate and list every individual network, you can list a group of networks together as one entry and this allows us to summarize the information so that we don't have as many entries. And then the third one, which was the other aspects, we've identified some strong security mechanisms and we have a field in the packet header which will allow future use for flow control, quality of service and metrics of that type. JUDGE SLOVITER: Could I ask a lay question, very basic? If you go to this four times as many bits is it going to increase four times everybody's address? THE WITNESS: The -- that's a very good question and actually that's something that a lot of people get confused. There are two ways that you look at addresses on the Internet. One is that bit pattern, currently the 32 bits, so the address of the computer sitting on my desk at Harvard is 128.103.65.15. Now, I don't expect you to remember that, I'm surprised that I do most of the time. JUDGE SLOVITER: And I don't intend to write to you that way cause I wouldn't know how. (Laughter.) THE WITNESS: Thank you. It's now in the record so you could look it up, but that's not the way that you should know about my computer. You should know about my computer by using its what is called domain name, which is a people friendly name, and that name is NEWDEV, as in the New Development Machine, dot Harvard dot EDU. As long as you're using that what is called the domain name, the size of the actual address, the number of bits in the address is not reflected back to something that the user has to deal with. I would not want to try to remember the 128 bit version of what my -- my computer's address is, but the domain name, NEWDEV.Harvard.EDU will remain the same. JUDGE SLOVITER: Thank you. Sorry. BY MR. BARON: Q Mr. Bradner, you also said on something called the IESG, the Internet Engineering Steering Group, correct? A That is correct. Q You've -- this is the Standards Approval Board of the IETF, correct? A That is correct. Q Take us through, very briefly, if you would, the standards track in terms of the three stages of standards, proposed, draft and full? A Actually, I would like to start a little bit before that. All documents which are going to be proposed for consideration for standardization within the IETF must first appear as one of the Internet drafts that you mentioned earlier that are publicly available ideas. And so somebody who wishes to, somebody or some working group or some group of individuals who wish to make a standard or have a document considered to be a -- for standardization creates a Internet draft. Usually that is the product of a working group or is that, you know, a working group is formed to look at that proposal, but not always. After working group consideration, the working group chair would propose to the area director within the area that this document be considered by the IESG for the standards process, for the standards track. The first step, the IESG then reviews that and does an internal vote and approves or does not approve of the document based on its technical quality, its clarity and all of the other things that one should consider when approving a standard. The first step in the standards process is, as you mentioned, the proposed standard status. A proposed standard is a document which is felt to be useful, i.e. has a constituency usually within a working group and within the IETF itself, and that constituency believes that this is of value to the community and that it has no known errors, no known flaws. If something is discovered in the process of evaluation by the IESG or the working group which is a flaw, then it should be returned to the working group and reworked. Six months after a document has been approved as a proposed standard it can then be considered for being a draft standard. To achieve draft standard status, a document, a specification must have multiple interoperable implementations, you know, it's got to be proven to work, and it's got to -- all of the individual aspects of it have to be proven to work, all of the individual functions have to have been shown to be implemented and interoperable. This is unlike some other standards bodies which just say this is a good idea and it's a better idea now than it used to be because we've thought about it longer. JUDGE SLOVITER: Could you give us an example? Bring us down to earth, give us an example of a standard like a proposed standard. What are we talking about? THE WITNESS: A proposed standard, for example -- JUDGE SLOVITER: Yes. THE WITNESS: -- an example of one is in the IE --in the Inter-- in the IET -- IP Next Generation there is a proposed standard which is the basic packet format and how that packet is handled by routers as it goes through the network. It defines the fields in the packet, the 128-bit addresses, what routers do when they encounter this packet, how they process it, all of that kind of aspect, all of those aspects surrounding defining a packet of IP Version Six IP Next Generation and how to move it through the network are part of a proposed standard. Another proposed standard would be -- well, a full standard is Telnet which is a, Telnet is the remote access protocol where you -- I can sit at this machine here and log into as if I were local to my computer sitting back on my desk at Harvard and that's a standard. So a proposed standard is: we think this is a good idea, we don't see any problems with it; draft standard is people have implemented it and it works and we don't see any problems with it still and more than one is implemented and they interoperate. And then four months after a document has been approved as a proposed standard, it can then be considered for full standard and full standard has to have the same implementation rules but it also has to be proven that people want to use it so that there is significant deployment. So we don't make something a full standard unless people are going to use it. BY MR. BARON: Q It is true, is it not, that apart from the IETF and the IESG, that there are other standards for the Internet that come from submissions by outside individuals or groups, correct? A There are, there are a number of bodies who make specifications, most of them call them specifications for some minutiae of legal ease that I don't quite understand, rather than standards. A number of -- JUDGE SLOVITER: Now you know how people feel if they don't quite understand when somebody says something. Go ahead. (Laughter.) THE WITNESS: I fully do understand, actually. I am a fish out of water here, so... There are many bodies who purport to make standards or specifications that are for use on the Internet. The IETF is the longest established of these and the one which has the most international flavor and the one which is the, well, I think anyway, since I'm a member of it, has the most credibility as an open forum for development of standards. We allow literally anybody who wants to participate. Many of the other groups have a membership mechanism where somebody purchases a membership or pays a membership fee and at the access to the standards either during development or when they're done are restricted, you have to pay for them. But there are a dozen or more different groups developing specifications for protocols to be used over the Internet, those groups are open, large consortia such as the W3C, the World Wide Web Consortium, or very focused ones such as the Master Card and Visa just announced a payments protocol to encrypt credit cards over the Net, and that is a small consortium and they have come up with a standard. And so there's a wide range of standards. Things in the Internet as in things in real life are standards only in the extent that people actually use what you've done. We can create something we say is a standard and if nobody uses it, well, we're whistling in the wind but it's not really a standard unless people use it. So the Web itself is something which did not develop out of the IETF standards process, it developed out of scientists wanting to avail the technology for use in over the Internet and this was -- this was developed outside the IETF, though now there's an activity within the IETF to codify and clarify the Web standards, the HTTP Standards. But, yes, there are many standards processes. BY MR. BARON: Q There are 53 or so full standards that have made it through this process? A Over the years, yes. Q These are common protocols -- A Something around that number. Q These are common protocols in widespread use on the Internet, correct? A They were at the time they were adopted. Not all of them are still in widespread use, some of them are quite historic. Q And there are some two dozen draft standards in the works, correct? A Somewhere around that number, yes. Q And about two or three more dozen proposed standards, correct? A That is correct. MR. BARON: Your Honors, I'm going to, with the Court's indulgence, approach the witness and provide him with an exhibit. We have provided -- JUDGE DALZELL: It's in our binder? MR. BARON: They're in black binders. JUDGE DALZELL: In the black binder. MR. BARON: In the Defendant's Exhibits 1 through 45, for the Court. I will hand the witness a volume as well. JUDGE SLOVITER: And this is going to illustrate everything he just said in black and white? (Laughter.) THE WITNESS: In a little bit more detail, I think. JUDGE DALZELL: Which exhibit is this again? BY MR. BARON: Q I wanted to turn to Exhibit 6, Mr. Bradner, I wanted to give a concrete example of something that the IETF is working on. You're familiar with this document, Mr. Bradner? A Make sure we're on the same page. This is -- Q It's -- A -- the charter for the address auto configuration working group? Q That's correct, marked as Defendant's Exhibit 6? A Yes. Q Your name appears on the first page of the document, correct? A That is correct. Q Could you explain how this document which is with title "Address Auto Configuration" will help unsophisticated computer purchases -- purchasers like myself to essentially plug and play when they buy computers? A The document itself won't help you a great deal. Q Okay. A But the -- this is, the document is a charter for a working group within the IETF, within the IP Next Generation area which is designed for to allow computers when they're taken out of a shipping carton and plugged into the wall to come up with that globally unique 128-bit address so that you don't have to type it in. You thought that remembering was bad, defining the right one and typing it was going to be awful. So this is a mechanism by which the computer can figure out a globally useful unique address and work with other technologies, particularly what is called Dynamic Host Configuration Protocol which is a way where a central administrator can control what address some particular computer gets. This is one of the activities of the IP Next Generation area. Q If I could re-formulate that, in other words, an individual does not have to obtain an IP address from some central source like Internet but an auto configuration will assign a globally unique address, correct? A It will assign a globally unique address but within constraints of and a range of addresses which has been provided from some central source, either directly or indirectly. It doesn't just go pick a number out of the air, it says that this network, this physical wire can have addresses 1 through 99 within this sub, sub-grouping of addresses and it will pick the one within that sub-group which uniquely identifies this machine but it does not affect what is called the high order bits or the more -- the more general part of the address which is supplied to it from a router on the local network, a computer on the local network. Q Am I correct in saying that each IP address is unique on earth? A That is incorrect. Q Let me, maybe I misphrased something from your deposition in the last week, let me quote from Page 55, Line 18. I'd be happy to supply the witness with a copy of the deposition transcript. A You can read it and -- JUDGE SLOVITER: Would you like the written deposition? THE WITNESS: Well, why doesn't he read it, if I'm still confused then I'll ask for a copy. BY MR. BARON: Q I asked a question that went: Question: "In lay person's terms it would mean that a person such as myself who may have difficulty loading in software or loading in whatever is required to put a computer -- to get it to go would have an easier time." You answered at some length, but at one paragraph you said "It will negotiate over the network for an address automatically and" -- here's the key section -- "assign an address which is globally unique and will uniquely locate this computer on the global Internet." Did I misstate the point? A Your question was whether every IP address in the world is unique and the answer is no. The -- the answer to the question on the address auto configuration is if the address auto, the node which is being configured is part of a network, part of a network which is directly connected on the Internet, then, yes, it will come up with a globally unique address. But there are very many, thousands and thousands of networks which are not connected to the global Internet and they are using addresses which may be the same as somebody else on the global Internet but it doesn't make any difference because they're not part of the same picture. And then there's a whole 'nother class which is getting increasingly common where an organization such as a university or a corporation, more likely a corporation, picks addresses which are convenient to it and then has what is called a fire wall between itself and the rest of the Internet and that fire wall translates the addresses which are local within its own corporation to addresses which are unique on the internet, but it does so not one address per node within the corporation but one address per speaker. So if I am -- want to just talk within the corporation, I never get an address which is unique on the Internet. If I want to go out and make a connection out on the Internet then I will be assigned the next address in the row of the ones available that are unique on the Internet. Normally there are very much fewer addresses on the window of the Internet than there are inside the corporation and the addresses on the window are reassigned by dynamically every time somebody connects and disconnects, makes a section through this fire wall and disconnects. This is because of the pressure of addresses on the Internet, we are still in a situation where the 32 bit IP Version Four address is under some stress in terms of availability, so in order to make it easier on corporations which may have very large internal networks but may not be able to obtain an address, a globally unique address, routable, globally unique address for every one of their internal nodes, they get a small subsection, maybe 500, maybe a thousand addresses which are reachable all over the Internet to deal with the 100,000 internal computers. It just means they can only have a thousand communications going on at once. JUDGE SLOVITER: Let me ask, could I follow that through? I had a question on that as I read his original affidavit. Let me give you an example and see if it has any relevance to this. The Federal Courts, the whole Federal Court system is in this circuit interconnected on what we call CC Mail and is in the process of becoming interconnected with Federal Courts throughout the country. But it's not currently on the Internet for various reasons, although there may be, I believe, several Internet addresses or -- I'm not sure that's the right technical way to put it -- that Courts or libraries within the Federal Court system are getting so that they can get the information available generally without compromising the security of the Federal Court communications. Now, is there a fire wall between -- is fire wall the right term that insulates the Federal Courts from the rest of the Internet? THE WITNESS: It could be and isn't necessarily, it could be simply that you have E-mail gateways such that Electronic Mail, and CC Mail is a product name, by the way. JUDGE SLOVITER: Mm-hmm. THE WITNESS: That Electronic Mail and CC Mail goes to a computer which then reconverts it from the CC Mail format into the Internet format so it can be forwarded out on the Internet. And then E-mail from the outside world can be reformatted and forwarded inside, without having the ability, for example, as I said Telnet, I could Telnet from here to my desk at Harvard. Without have-- the E-mail gateway would not permit the passage of Telnet packets so that somebody from outside couldn't try and connect up and use one of the internal court machines. So there are different ways to get that isolation. Fire Walls is one of them, Application Gateways is another one. The modern Fire Walls tend to be Application Gateways built into a single box, a number of different application gateways, a Telnet Gateway, an FDP Gateway, a Web Gateway, and an E-mail Gateway all built into the same box and many of them do this address translation. In your case it's more likely, speaking just as a general indication, that the addresses inside are not even translated to addresses outside, that the message is received by the Gateway and retransmitted as if it were an entirely new message using the address of the Gateway when it's going out on the Internet, nothing related to the individual source node where the message came from. JUDGE SLOVITER: And there would be then, is it correct that if you use such a gateway or whatever the communication process is, there would be no way outside to know what's really -- what's coming in inside or where it's going inside? THE WITNESS: That -- that actually is a key point. We don't -- the Fire-- one of the aspects of Fire Walls is to try and protect the knowledge of the structure of the inside network from the outside, it's to hide the inside structure. So from the outside, if I have an -- if I had your E-mail address, I could send you E-mail but I wouldn't know how that would get to you once it got past this gateway. I wouldn't know, wouldn't be able to determine from the outside anything to do with the structure of the Court network nor what computers were there, where you read your E-mail or anything. It's one of the functions of Gateways is to protect the internal structure from visibility. JUDGE SLOVITER: So then it has at least two objects. One is because there are a multiplying number of addresses and there are just or may not be enough addresses and the other or many others are for other purposes? THE WITNESS: Yes. JUDGE SLOVITER: Okay. BY MR. BARON: Q Just a couple more questions on standards. You would agree that a number of organizations are responsible for the development of communications and operational standards and protocols used on the Internet, correct? A A number of organizations believe they are, yes. Q The Internet wouldn't exist today as we know it without some standards or some rules of the road, correct? A That is correct. Q And you recall saying in your deposition to me that we are in a, quote, "standards development rich environment," unquote, on the Internet and you stand by that? A Yes, or sit by it or whatever. Q You stated in your supplemental declaration that you have, quote, "A complete understanding of how communications are accomplished on the Internet today, including communications such as E-mail use, Net and World Wide Web," correct? Why don't we break down the Internet and start with World Wide Web since most of the plaintiffs in this case in the lawsuit have Web pages. Mr. Bradner, can you describe for the Court what the World Wide Web is? A The World Wide Web is basically two things: it's the URL that you mentioned earlier which is a pointer, a way to -- a way to identify a particular location and piece of information within that location on the Net and software that interprets those pointers and goes off and retrieves the documents that's been referenced by the URL. Q You testified at your deposition last Friday and I'm paraphrasing this, but correct me if I misstate something, that the World Wide Web is a concept more than anything else, it is comprised of a number of servers which can provide information about requests in the same general concept as other servers, new servers, FTP servers and the like and a descriptive language which allows you to embed in a piece of text locators defined to point to other documents. Is that a good statement? A The World Wide Web uses a -- JUDGE DALZELL: You have to say yes or no so they can get that. THE WITNESS: Oh, yes, sorry. Yes. MR. BARON: Thank you. BY MR. BARON: Q The World Wide Web uses a graphical user interface, correct? A The -- the World Wide Web client applications that I have seen use a graphical user interface. Q Why don't you describe for the Court what a graphical user interface is? A The early computer interfaces tended to be character lined, lined character type interfaces where you typed words and commands like if you've used DOS, it's a DOS interface, it's where your view of the Net or your view of the command into the computer is one which is a character stream, you type in words with varying degrees of meaningfulness and asking it to do something. A graphical user interface tends to be a full screen application where you have a -- an ability to, with a mouse or with cursor, those little arrow keys on the keyboard, locate something on the screen and tell it to activate a program or to fetch a file or do something because you're selecting something on the keyboard -- something on the screen, rather than typing the name of something in on the keyboard. Q The graphical user interface was designed to be user friendly, correct? A The hope of the designers of graphical user interfaces is that they're user friendly. The definition of "user" and "friendly" are to the mind of the beholder. (Laughter.) Q In fact, the Web's user interface was designed to allow people with a wide variety of computer skills, indeed even with some -- some with minimal computer skills to access vast quantities of information, correct? A That is correct. Q And the language for creating Web pages on the World Wide Web was designed in a way that makes pages easy to write, makes it easy to put up pages on servers and makes it easy to distribute information around the world, correct? A That is -- that was the statement of the people who designed the language but I do notice that many of the books on HTTP which is this language tend to be in the one to two-inch or three-inch thick variety. So again this might, it somewhat depends on one's interpretation of the word "easy." Q HTTP or HTML? A Oh, HTML, sorry. Right. I get those -- Q You stated last week in your deposition -- JUDGE DALZELL: HTML. MR. BARON: That was going to be my next question, your Honor. BY MR. BARON: Q Why don't you state for the Court what HTML is? A Hypertext Markup Language or something of that general ilk. There's -- Q Were -- A There's too many acronyms in this business. JUDGE DALZELL: May I be the first to agree with you? (Laughter.) JUDGE SLOVITER: And I'll be the second. MR. BARON: We're going to get to HTML, Judge Dalzell. BY MR. BARON: Q You stated last week in your deposition that you've looked at thousands of Web pages and that there are probably tens of thousands of Web pages in existence. That's correct, right? A Well, the tens of thousands would be a -- what I meant in when we're speaking of that is there are tens of thousands of locations where Web pages exist. The actual number of Web pages in the sense of a screen image that you could retrieve is certainly in the millions. I know I have on my own site which is one server, one Web server with one home page, there are thousands of screens that you can retrieve. So if you're talking about Web pages in terms of images on a screen, then there are millions of them. Q Okay. Now, apart from individuals -- JUDGE DALZELL: Excuse me, you say you have a Web page? THE WITNESS: I have a Web server. One of the things that I do at Harvard is to run a test lab which examines the performance of routers and things like, network devices like that. And I put all of the information that I've gotten from this examination up on line for anybody to take a look at and there is thousands of pages, mostly of numbers and some of pictures of performance curves available from the Web server which is running on the computer sitting on my desk. BY MR. BARON: Q Apart from individuals, Mr. Bradner, it would be a fair statement to say that organizations including commercial organizations such as companies selling potato chips or pencils or cars use the Web as a way to provide information, correct? A Correct. Q And to sell their products, correct? A At this point, more to provide information. In the future, in the near future I trust, more will be in the business of selling their products over the Internet. Right now because of concerns of security and things of that nature, few, relatively few companies are actually doing retail over the Net, going and buying a bag of potato chips over the Net is not something that is readily available today although you can order a pizza if you happen to live in Santa Clara, California. Delivery is a big problem if you're doing it from here, but -- (Laughter.) A -- that -- soon you'll be able to do that. So I'm just nuancing on the word you use of selling because right at the moment it's more providing information than it is selling. JUDGE DALZELL: Because you can't close the sale? THE WITNESS: Actually you can and you can by putting your credit card number in and actually the credit card transaction over the Internet today is more secure than giving your credit card to the waitress at the local restaurant, but there is a feeling that it is not as secure. And so there aren't many -- JUDGE DALZELL: I thought there was a problem of verification? THE WITNESS: The -- it's the same -- JUDGE DALZELL: Or so we're told. THE WITNESS: Well, it's the same level of problem of verification as what happens when someone calls up and orders something from L.L. Bean over the telephone. L.L. Bean has to go through a process with which they call up the credit card company and say is this a valid credit card. JUDGE DALZELL: And is it not true that you still have to go outside the Internet to do that process? THE WITNESS: Today that is true. I would hope that in the relatively near future -- JUDGE DALZELL: Defined as? THE WITNESS: Well, Master Card and Visa did define a language for moving of information about credit cards over the Net, they said it would be, that this definition would be available I think this month or next month. So in the next six to nine months the function set to be able to send a secure credit card to Master Card to ask whether it's a valid card and whether the person has enough money to pay two dollars or whatever your fee is going to be should be there, but this is a projection rather than a statement from knowledge of who is developing these. JUDGE SLOVITER: How could phoning tell or assure that it's a valid card? It might show or how can phoning assure that X, that Judge Dalzell who gives the card number is in fact Judge Dalzell rather than Judge Buckwalter? THE WITNESS: Actually, it does not. And that's --it does not now when you call up for one of these mail order houses. They do it on a basis generally of two things. One is that in general when you order something you order it, particularly if it's going to be shipped to you, you order it shipped to you so that in some cases like American Express, if it's a valuable shipment, will verify it's being shipped to the billing address. And if it's not being shipped to the billing address, you have to call them up and tell them no, this is a special case and I want it to go someplace else. I know because I had to do that. Other credit card just ship -- know because they have the shipping address of where it's going to, they have an audit trail so in case somebody protests that this wasn't, it wasn't me who placed this order, they can then do some kind of tracking to try and figure out who it was who placed the order. And the same thing would be true over the Internet. JUDGE SLOVITER: But would that be time consuming or now that we have computers could that be easy and instantaneous? THE WITNESS: Well, in a real way ordering something over the Internet over ordering something from a mail order house over the telephone isn't going to change any of the mechanisms involved other than how do you do it. You sit there with a Web page and do some clicking on with your mouse versus you call up on the telephone and tell the nice person who answers the phone that you want an item on Page 67. The rest of this, what happens behind the scenes, works the same way today. There will be an increase in efficiency when the verification process for verifying the card, instead of requiring a separate communication normally by a phone line with Master Card or Visa or American Express could be done with electronic communication over the Internet. That will be a change in efficiency, but it doesn't change the basic functionality which is they're depending on you or your knowledge of the credit card number as your identifier to identify yourself and the fact that they can trace where the order was sent to as sort of a second guess to figure out what happened when something goes awry. BY MR. BARON: Q Putting aside actually ordering merchandise by use of a verified credit card via the Internet, it's certainly true, isn't it the case that both individuals and companies can have Web pages and that have a phone number on them or an 800 number or a toll free number for people to call to buy things that they see advertised on the Net, isn't that correct? A Yes, that is true. Q Okay. Before we get into what individuals and nonprofit organizations other than corporations can or cannot do, let's talk about some technical matters including some -- JUDGE SLOVITER: You mean we haven't been? (Laughter.) MR. BARON: Some more technical matters. BY MR. BARON: Q Including some descriptive language used for the World Wide Web. Mr. Bradner, could you tell the Court what a Web server is? A A Web server, a server in general in computer jargon is software which is running on a computer which is waiting patiently for a command to be sent to it over a network and that command, if it's an FTP server, it would be an FTP command, FTP stands for File Transfer Protocol. If it's a server which returns phone numbers it's going to be a phone number query. If it's a database server, so lots of different servers, they have the same basic function which is just software running in the computer waiting for a query. A Web server is one which is waiting for a query which is in -- over the Net which is formed in Webese, in the right format for a Web query. Q You told me last week that the World Wide Web is sewn together with URL's, is that a fair statement? A Yes. Sorry. Q Now, on a particular Web page there can be pointers to other pages on the Web, correct? A Those are the URL's of which we were just speaking. Q And the pointers -- all right, they can be pointers to other URL's. And Web pages can also have -- A The pointers are the URL's. Q Oh, the pointers are the URL's, okay, I stand corrected. Web pages can also have pointers to files which contain audio or sound, correct? A That is correct. Q In fact, Web pages can contain pointers to files in any one of a number of forms containing any one of a number of things such as text, sounds, still graphics or motion graphics, correct? A That is correct. Q One can take a home movie on a Camcorder and digitize it and transpose it in a way that would be viewable by clicking on a pointer on a Web page, correct? A Assuming that the person who had the client who had the Web browser had the right software installed which allowed them to download and then view motion graphics and assuming that the motion graphics were stored in a format compatible with the browser that the individual had. Both of those are not assumptions you can make a hundred percent, but still given that qualification, yes. Q Could you tell the Court what a browser is? A A browser is the jargon term for a Web client. The client is the software running on a user's computer to access some server and a Web browser is the software running on the user's computer to access a Web server. Q And what is a search engine? A A search engine in this -- in the context of the Web is a piece of software which, when given a query, it's a database query responder, it's a server for database queries, it --you give it some information about something that you wish to find and it goes to its database and tries to find it in that database. Search engines have fine degrees of sophistication of ability to take just single words or words in context or concepts in the sense of you can give some search engines a piece of text, a newspaper article, and say this is interesting to me, find other things that look like this. And it's quite -- some of them are very sophisticated. They look at their internal database to try and find other things, other references in that database which are compatible with the query that you gave it. Q Let's get to the heart of things, Mr. Bradner, by discussing something called HTML. JUDGE SLOVITER: Before we do that maybe we should let the witness have a break and we should all have a break. Okay? JUDGE DALZELL: I agree. JUDGE SLOVITER: Ten minutes, I'm told. THE COURT CLERK: Please rise. (Court in recess; 10:40 to 10:55 o'clock a.m.) JUDGE DALZELL: All right, Mr. Baron. MR. BARON: Excuse me, your Honor, we were just taking care of some housekeeping functions. (Pause.) BY MR. BARON: Q Mr. Bradner, we were about to discuss HTML, could you tell the Court what HTML is? A It's a language, a descriptor language which is used to define within a Web server how a document should appear on the screen of the Web client, the browser. Q Perhaps an example of HTML code would be helpful here. Could you turn to Defendant's Exhibit that's marked 14 in the black binder? (Pause.) Q Do you have that? A Yes, I do. Q Mr. Bradner, does this appear to you as the same exhibit that I showed you at your deposition last Friday? A Yes. Q This represents the Worldwide Web home page of an organization entitled Stop Prisoner Rape, which is one of the plaintiffs in this lawsuit. And you will note -- and you would agree, would you not, Mr. Bradner, that the first four pages represent Web pages in their usual format and behind those four pages is a series of pages which represent the same text but in HTML code format, is that correct? A That appears -- that is what it appears to be, yes. Q Looking at the immediate page behind the usual format Web pages, the top of the page says -- JUDGE SLOVITER: These aren't paginated, are they? MR. BARON: No, they are not, your Honor. (Discussion held off the record.) JUDGE DALZELL: You're talking about the first page following the conventional -- MR. BARON: That's correct, the -- JUDGE DALZELL: -- conventionally arranged text? JUDGE SLOVITER: So, the one that says -- MR. BARON: That's correct and the -- JUDGE SLOVITER: That's right, okay. MR. BARON: That's correct, your Honor. BY MR. BARON: Q You see the bracket HTML and bracket Head, correct, Mr. Bradner? A Yes. Q The designation Head represents the head of this HTML document, correct? A Yes. Q And you see the term Meta in the third and the fifth line? A Yes. Q What does the Meta represent? A As I said in my deposition and when we talked last Friday, I did not and do not represent myself as an expert in HTML. So, I would suggest that if you want to investigate the details of HTML it would probably be better to ask somebody who is. Q But looking at it you're certainly more expert than I, that the key words here are words that are in a field in a Meta tag in the header, correct? A Yes. Q Okay. And there's a body to an HTML document, correct? A Yes. Q And down at the bottom of this page there is a reference to a URL. The HTML source code includes references to particular URL's as a usual course, correct? A Most of them do, yes. JUDGE SLOVITER: Oh -- JUDGE DALZELL: The very bottom, the very bottom. MR. BARON: The very bottom of the page, your Honor, it says "Bracket A-HREF equals," and then A-URL, which represents another Web site. THE WITNESS: Actually all you can tell about that, URL, is that it represents a particular document someplace which may or may not be on another site. JUDGE SLOVITER: Where does it says URL? JUDGE DALZELL: He said that is the URL. JUDGE SLOVITER: Oh, okay. BY MR. BARON: Q Yes, that's a better description, Mr. Bradner. Now, in your deposition last week you indicated that the type of parental control rating scheme you preferred would be one in which an individual's browser could be configured to send a copy of a particular URL, including a URL in a document, to a third-party rating service with a query to the rating service, asking for information about the contents of the URL, correct? A Actually, to be very precise, about the contents of the file or document pointed to by the URL. Q Okay. Now, this would be one of the methodologies suggested by the PICS scheme, P-I-C-S, which is a parental control rating scheme being worked on by the W-3 consortium located at MIT, correct? A That's correct. Q And that's the scheme that's embodied in Defendant's Exhibit 15, if you could turn to that, the document which says, "PICS: Internet access controls without censorship"? A Yes, this is a document you showed me last week. Q It is true, is it not, Mr. Bradner, that a browser under this model of parental controls could look to the specific header information in HTML source code for a tag or a label that's put in the header by the content provider as part of the overall rating scheme, isn't that correct? A To be clear, you had just asked me about my preference for a third-party rating service, it appears that you're asking me now about PICS as a general concept, I just want to be sure what it is that you're asking me. Q Well, I'm asking PICS as a general concept. A Okay. So, in PICS as a general concept you -- PICS defines tags that you can place into a document, into the header of a document, HTML document and other documents, which can be used to convey information about the -- some content of this document, that is correct. Q And on Exhibit 15 at Page 6 of 9, at the bottom left-hand corner, that's where the pages are identified, the second full paragraph, if you would read along with me -- JUDGE BUCKWALTER: Well, where are you? JUDGE DALZELL: Page 6 of 9. MR. BARON: It's Exhibit 15 and it's Page 6 of 9, you can see at the bottom left hand of the document. BY MR. BARON: Q I'm going to concentrate on the second full paragraph, starting with "Since," the word, "Since." And the second sentence says, "The first is to" -- the first methodology of PICS, is that correct, Mr. Bradner? A That's actually a third sentence. Q Well, it says, "The first is to embed labels in HTML documents. This method will be helpful for those who wish to label content they have created." That's one of the methodologies embodied in the PICS parental control rating standard, correct? A That's one of the -- that's one of the methods in their proposal, yes. Q Okay. Indeed, you believe that as a technical matter one can embed a character stream which could be interpreted by browsers or other software if it is so desired, correct? A In certain documents, certain types of files and documents that is correct, in other types of files and documents it's incorrect. Q Well, it's your view, is it not, Mr. Bradner, that as a technical matter of ease or difficulty that it is trivial to embed a tag or a label in HTML source code, correct? A It's a matter of typing a few characters, so, yes, in concept; in implementation, if you have thousands of pages of source code then it might be a little difficult, but in concept it's easy, you just type in the character string. Q You told me last Friday in your deposition that for your own Web site, your own Web pages, the home page it would be trivial to embed a tag, you could do it in five minutes, correct? A Well, actually it's a little more -- I said a little bit more than that. My current Web server I do not happen to have a document which is a home page. The Web server points to a part of my -- the directory tree in my computer and it has automatically created a home page, because I haven't gotten around to creating one myself. So, it would take more than five minutes, because I would have to create the document in which to embed the string before embedding the string and I couldn't tell how long that would take, it would depend on how anal I got and how pretty a picture I wanted on it. Q Well, you said at Page 223 of your deposition -- MR. BARON: -- I'd be happy at any appropriate point to hand the witness the deposition if it will be -- JUDGE DALZELL: Whenever you want it, you just say so. MR. BARON: -- helpful for the record. BY MR. BARON: Q You said at Page 223, Line 8, "Certainly on my site it would be trivial for me to do," correct? A Once I created a home page it would be trivial for me to do it, yes. Q Could you turn to Defendant's Exhibit 16? This exhibit is one that I showed you last week, correct? A That's correct. Q It's titled, "Safe Surf Internet Rating Standard," are you familiar with Safe Surf? A As to the extent that you showed it to me last week. Q Okay. On the second page of this exhibit at the top, the first full sentence says, "If a majority of them spent five to ten minutes to implement the system by marking their site then a child-safe Internet could be realized in a matter of weeks." Do you see that statement? A I see that statement. Q And do you agree with it? A No. Q You could build PICS compatible software into existing browsers, correct? A One could, I wouldn't proclaim to be a good enough programmer to in any particular case. Q That's technically feasible, correct? A That's correct. Q Back to Exhibit 15, looking at the bottom of the Page 5 of 9 in the document. It's the page with blue Figure 4 at the top, but I'm going to concentrate on the bottom of the page. Do you see the sentence that starts, "Anything"? A Yes, I do. Q Let me read it to the record: "Anything that can be named by a URL can be labeled, including resources that are accessed via FTP, Gopher or Net News, as well as HTTP." You agree, do you not, Mr. Bradner, that you may extend URL's to provide labeling in some form across these applications on the Internet, correct? A To be very specific and concrete, you can extend the format of URL's themselves to include additional information, which could be used by a browser to decide on whether to implement -- to instigate a particular application. You would not actually do anything in the application itself, for example FTP, you wouldn't modify FTP, you would modify the browser to decide on whether or not to start up FTP based on additional information in the URL. JUDGE DALZELL: I want to get very concrete on this, because it's an important issue. The Carnegie Library, Mr. Croneberger is here for the Carnegie Library, it's card catalogue is on line. Now, I take it the card catalogue is a site, correct, it has a URL -- if I want to get to it it has a URL, does it not? THE WITNESS: The -- I can speak with knowledge about the Harvard University College -- JUDGE DALZELL: All right, fine, take that. THE WITNESS: -- Library. The Harvard University College Library, which is called Holis (ph.), is available as an interactive program. So that you would Telnet to a server at Harvard and then it presents a screen wherein you can do an author search or a title search or things like that. JUDGE DALZELL: Well, what I'm getting at is what I think Mr. Barn is asking you, is the idea here that Harvard or the Carnegie Library would rate its card catalogue? THE WITNESS: In the context of Harvard's, Harvard's Holis system, what would have to happen is any place where somebody referenced Harvard's Holis system, a URL which referenced it, any place where that any URL existed the reference Holis would have to be extended to include a rating of Harvard's system. JUDGE DALZELL: That's what I mean. THE WITNESS: This wouldn't be Harvard rating it, because Harvard isn't creating the URL's that might be placed at Brown or at the National Library Association or any place else, because that is a pointer to Harvard and it's the pointer in this concept which is modified, not the site itself. In this particular case you do not get to Harvard -- the way you access Harvard doesn't give an interactive way for a browser to ask Harvard what its PICS rating is. JUDGE DALZELL: Well, then I'm not understanding this at all. The PICS rating -- assume that everybody adopts this PICS system, okay? Will the Harvard card catalogue that's on line, will it be rated or will just subsets of it be rated? THE WITNESS: There -- JUDGE SLOVITER: And who has to do that rating? JUDGE DALZELL: Exactly, and who has to do the rating? THE WITNESS: All right. I think using the Harvard catalogue is exactly the kind of case where we can look at it. The current technology, the current way the Harvard catalogue is implemented is that you interact with the Harvard catalogue with the same program that I would use to interact with my computer sitting on my desk, which is Telnet, it allows you to remotely be connected to that computer and remotely interact with that computer as if you were a local terminal; this is not a Web interface, it is a local terminal interface. In that context Harvard has no way of rating -- have no way of handing back a rating to anybody, because what would have to happen instead is -- it's like you would put ratings in T.V. Guide of T.V. shows, it's not that the ratings are embedded in the shows, it's every place where somebody pointed at Holis you would have to have that place which did the pointing have the rating in it. So, Harvard wouldn't have control over that. JUDGE DALZELL: Yeah, but what I think Mr. Baron is getting at is the feasibility of if you are going through a card catalogue on line, which Mr. Croneberger describes in detail in his declaration, would this marker be right next to "Rebecca of Sullybrook Farm," and that's G rated, but then when it has an Ice T lyric it would be NC-17? THE WITNESS: Again -- JUDGE DALZELL: I'm not being facetious here. THE WITNESS: No, no, I agree. Under the current Harvard system we wouldn't be able to implement this, I'm saying under the current Harvard system that the pointers are outside of Harvard's jurisdiction because they're pointers to Harvard, not pointers within Harvard. So, other people would rate Harvard. Another interface to this -- the same facility which does not currently exist, but could be made to exist, would be a Web Browser-type of interface to the Harvard College library system. In that case the browser could be able to see a rating and the rating would be actually buried in the URL, when you said Ice T the URL, which specified where the file was if you're going to retrieve it, then that URL, you could embed in that URL the PICS parental warning symbol. JUDGE DALZELL: But my point, and it's a very important point to this case, is since we know at least at the Carnegie Library, and I would think that's in the Carnegie Library would be up at Harvard, that they have the Ice T lyrics, is the whole card catalogue NC-17, to take the MPAA rating -- THE WITNESS: In the -- JUDGE DALZELL: -- because there is some dirty words there, in some people's view? THE WITNESS: The question that I was asked a little while ago, whether it was easy -- the statement in the exhibit here of whether it was easy for everybody just to do this, assumes the very assumption -- the question you just asked, which is that, yes, Harvard would have to rate its entire catalogue as questionable because of some references within that catalogue. The effort to go through and rate every individual reference within the catalogue -- Harvard's -- Carnegie Mellon's catalogue is a subset of Harvard's, of course -- (Laughter.) THE WITNESS: -- it's some six or seven million references in the Harvard catalogue, though I think on line is three or four million at the moment, this would take considerable effort to go through and -- JUDGE DALZELL: And rate them. THE WITNESS: -- rate every single one of them. JUDGE SLOVITER: If we started -- if nobody had ever put Shakespeare into this -- ever at all put it into the system and somebody, a third party or somebody else went through Shakespeare before they did this and began to rate Shakespeare plays, is there some feasible method where anybody, any library that then have Shakespeare could absorb that rating? Or if Judge Dalzell, who has a younger person, unlike mine, who can read anything, but would he be able to find some mechanism whether she or he, I don't know, looked at Shakespeare, wherever it might be? THE WITNESS: There's two aspects to that and actually something I should clarify. On the Harvard College Library this is the library card catalogue, not the materials itself, there are other libraries with materials itself on line. For example, I was researching for a column that I do and I was looking up Flatland, which is a -- some of you may have read that, it's from the late 1800's, it's about a world of two dimensions -- JUDGE DALZELL: I read it in geometry. THE WITNESS: Yes, well, you should read it at least there. And I wanted to look at it, because I was going to do a column which happened to be based on that. So, I did a Web search and I came up with a site where the text for that book was on line, and I went on off and I read it. And this was a library which provided this, it's one of the university libraries, I forget which one, where that material was on line. And I think the questions you were asking are more related to places where the material is on line -- JUDGE SLOVITER: Exactly. THE WITNESS: -- rather than the Harvard University catalogue, which is just saying, well, the rap songs are available by going to the stacks and looking in Bin 3. So, in the areas where the material is on line that's a much more complex issue, that -- certainly the Harvard -- I don't think the Harvard catalogue, the catalogue per se would be ever considered verboten, but certainly some of the items within that the catalogue references could be. There is a mechanism where one in theory could do this. A lot of college libraries, a lot of libraries in generally actually use external sources when they create their card catalogue, they send a list of titles to a commercial firm which has expanded information about titles. So, you send -- you say I've got Shakespeare's Hamlet and Edition 14, give a little bit more information, they return to you the information block, which includes the key words for use in searches and all of the other information that you might want for your on-line reference to this document, rather than you having to enter all of this -- the individual university library having to enter all of this information they go off to this third party. And in theory that third party, if the rating has --if a rating has been done that third party could include that rating in that block of information that they return when the university or other library says tell me about Shakespeare's Hamlet, 1912 Edition from whatever. BY MR. BARON: Q Well, I'd like you to return, however, to the methodology that I pointed you to in Defendant's Exhibit 15, which is that one of the methodologies in PICS, is it not, to -- that the content creator, the content provider embed the tag in their document rather than a third-party rating organization, correct? A As I said before, that is a feasible and reasonable thing to do for some document, it is not possible for others; it is not possible for binary files, for executables, for example, you can't embed something in there because it would destroy the integrity of the file itself -- Q All right, but for -- A -- it would make the file itself useless. Q But for the Web pages that represent, for example, the Stop Prisoner Rape Web page, that doesn't have a binary or an executable file, so far as you know? A As far as I know. It could put at the top of the page -- embed in the HTML a coding, that is correct. Q It is also technically feasible to tag a portion of a Web site, correct? A There is no -- in the Web there is nothing which -- there is no structure which says this is a portion of a site and this is not. Going back to your question earlier about the URL that was at the bottom of the page, I made the point of saying that this was a pointer to a file some place on some server, there is nothing to say that this is structurally on this server or any other server. So that if you are -- if all of the access to some subsection of your disk is through a particular home page and there are no URL's that exist any place else in the world which have a more explicit pointing down inside of a sub-subdirectory then, yes, if you put some kind of labeling on the home page, on the first page of this sub-tree you could imply something about the rest of the tree. But that would only be making, again, the assumption that nobody had a URL which pointed further down into that tree, if they did they would never even look at that page, they would go directly to the more specific document. Q Well, I'm at a point where I think it's reasonable to read a portion of your deposition last Friday and see if we can seek clarification here, I'm at Page 222. Let me read into the record -- MR. BARON: -- and, with the Courts' indulgence, I think it would be appropriate to show the witness the deposition. (Pause.) BY MR. BARON: Q I'm at Page 222 and around Line 17. MR. BARON: If the Court wishes, I have copies of some format of the deposition, but I do intend to read a few sentences here. BY MR. BARON: Q You're answering me and you say, the witness, this is at 222, Line 17: "I could make a label and I could see that most people could make a label and what, for a lack of a better term, home page for the site which in some way characterized the contents of the site and do that quite economically, yes. It gets a little more complicated, the site is like a library site that are flat laying on board, where the characterizations of the contents vary on a per-file basis." Going down to Line 8: "Certainly on my site it would be trivial for me to do, once I got the software and got everything else and got a sample page to put up it would probably take me five minutes to do that after I got all of the crap in line, a technical term." Moving on to Line 14: "And so, yes, it would be economically feasible if indeed somebody" -- MR. MORRIS: Your Honor, I would just ask that Mr. Baron read the entire page, he is leaving out some important points that -- and place the -- JUDGE DALZELL: Well, you can get that -- JUDGE SLOVITER: Well, you get cross-examination and the witness has got the entire testimony in front of him. MR. MORRIS: Okay, that's fine. BY MR. BARON: Q Continuing at Line 14: "And so, yes, it would be economically feasible if indeed somebody were to distribute a sample. Everything below here is fine file, putting that into my environment would be actually quite easy to do." Is that still your testimony, Mr. Bradner? A Yes, and that -- and that's absolutely true and I think that's what I just said. But it makes one assumption, which I did not state when we talked last week which I did just state, which is it makes the assumption that anybody referencing my site would only have a reference to my, quote, "home page," rather than a more explicitly reference to some subsection point, which actually in my particular case I know is not true. In my particular case some individual vendors of equipment provide URL's pointing to their results, which point down inside of my site, bypassing my home page and they are saying, go and look at this file, which is underneath this directory, underneath this directory, underneath this directory, and go look at the results there. So, yes it is true that I could modify a -- put in a home page, but that is only effective if people look at -- are actually stopping at the home page on the way to what they're looking for and that may or may not be true. Q The concept of coming up with some form of a standard way to tag or label a warning sign is perfectly reasonable, correct? A Yes. Q And it's technically possible, correct? A Yes. Q Assuming that there was software or browsers in the marketplace that could read the tag or label in HTML source code that Web site would be blocked, correct? A Again, it would be blocked if indeed that particular Web page was one that the browser referenced on its way to the document that it was seeking. In my case, I included in my news column a URL for Flatland and that URL specified the file which is Flatland's home page, not the file which is the home page of the library system itself. So, if I -- if somebody used the URL that I provided in the column they would bypass any home page of the entire site and would go directly to the Flatland file and would not see any tags that happen to a site-wide tag, because their browser would never read that page. Q You stated in your supplemental declaration filed on Tuesday at Paragraph 79 that, quote, "To my knowledge no Internet access software or Worldwide Web browsers are currently configurable to block material with such tags." Do you recall that statement? A Yes, I do. Q You stated in the deposition that, however, the Netscape owns the lion's share of the browser market, around 80 percent of the market, correct? A I think I stated that Netscape has stated that they own 80 percent of the market. (Laughter.) Q Last Friday -- JUDGE DALZELL: And they're not under oath. (Laughter.) BY MR. BARON: Q Last Friday at your deposition I asked you specifically how difficult would it be for Netscape to tweak its browser to understand a tag or a label embedded in a header in HTML that said adult, was in fact a site that was adult, and you responded that, quote, "I certainly don't think it would be an inordinate burden to do something of that form." You stand by that statement? A Yes. Q You also agreed as a matter of technical feasibility that Microsoft could do the same, correct? A Yes. Q And programs could be changed at AOL, Compuserve and Prodigy to do the same, correct? A Yes. Q And Surf Watch and Cyber Patrol and the world of that --of parental control software, they could change their software programs to pick up the tags or labels, correct? A They can pick -- they can -- software can be changed to pick up the labels whenever they examine a page that has labels in it. Q Okay, we're going to leave tags and labels. Let's turn to directories and registers in cyberspace, particularly on the Web. You recall at your deposition that I asked you whether you agreed with the statement that many people believe there should be a white pages directory for the Internet and you at least conceded that many people do believe that, correct? A Yes, that's correct, I conceded that. Q Even if a comprehensive index to the net is impractical in some sense you surely agree, do you not, Mr. Bradner, that a white pages subset of cyberspace is technically feasible, correct? A It's more than technically feasible, there are a number of organizations claiming they are providing just such a thing. Q In fact aren't there, as you said, many neutral places or sites that exist where URL's can be picked up in a kind of index or directory, correct? A Neutral and non-neutral, yes. Q Indeed, you told me last Friday that a URL is a URL is a URL and that no technical issues are involved in creating pages which list URL's, correct? A That is correct. That actually is the point I was making earlier, that if there is a URL pointing to a -- pointing to the Harvard College Library that -- and we're making the assumption that we're controlling access by putting PICS-type tagging in the URL's, it's wherever that URL exists, whether it's on Harvard- owned machines or anybody else's machine, which is where that labeling would have to be done. And if there is 10,000 places around the world which have URL's pointing to Harvard, all of those 10,000 places would have to rate -- would include the ratings for Harvard in their URL's, it would not be under Harvard's control to make them do such a thing. Q Aside from indexes or directories, if you have content -- if you are a content provider and you have content you wish to restrict, for whatever reason, you could call Surf Watch or other parental-control products to let them know about your site in cyberspace, correct? A We had a long discussion of this last Friday and the clear statement is yes, of course, I could call Surf Watch and do so, but Surf Watch would have very little way of knowing whether I had the authority to make a statement about a particular site, they would have to have some ability to resolve that this person had some relationship to the site that was being spoken of. So, if I'm a do-gooder and wanted to talk about some other site I may or may not have the authority or maybe I'm just trying to be mean to somebody or they are a competitor of mine, Surf Watch would have to go through some mechanism to insure that I had the right to speak of that site. So, in a true literal sense, sure, I could call up Surf Watch and say the Reuter (ph.) vendors think that the information about their products on my site is dirty because it paints them in a bad light and you should block that, I could do that, but I would suspect Surf Watch would be a little curious as to why -- whether I had the right to do that. Q Well, I'm concentrating on the good-faith actions of content providers and you have conceded that they could certainly call Surf Watch -- you can E-mail Surf Watch, right? A I don't know their E-mail address, but I assume that they're on the net -- Q You could fax -- A -- it would be silly if they were not. Q You could fax Surf Watch? A Again, I don't know their fax number, but I assume you can. Q You could hyperlink Surf Watch from your site, correct? JUDGE SLOVITER: What would this -- let's get back to the question and what would it say, if you faxed Surf Watch what is your question, so we can -- MR. BARON: The question is whether a content provider could take an affirmative action if they had a site that they wished to block because of whatever reason, for example, that it was not appropriate for minors, and they wanted to inform the parental control software companies that are out there, and Surf Watch is my example, could they take an easy, simple action to E-mail, fax, telephone or hyperlink that parental control software company to let them know that the site in cyberspace exists, that's what these questions are. JUDGE DALZELL: And your answer is yes? THE WITNESS: Except for the last one, hyperlink, I'm not sure what he means by that. BY MR. BARON: Q You can just put a click on a Web site and click it to the Surf Watch and they would -- it would be a link to them. A That would bring up Surf Watch's home page, I'm not sure what that would gain us. Q All right. Well, putting that aside, wouldn't doing any of these affirmative actions cure the reliability problems that you yourself have stated with respect to Surf Watch? A the reliability problems I believe that you're referring to are where I said that there was a window of vulnerability, if a primary method by which one of these blockers is working is that you have a list of sites which is distributed at some periodicity to update the local copy of the browser, there is a window between the time that a site comes on line and the time the site is discovered and the time that this update occurs, there's a window of vulnerability wherein Surf Watch wouldn't block a site that it otherwise would. And if indeed there were some reliable methodology for getting a message to Surf Watch indicating that this site is a funny site, and I'm in control of this site and I tell you it's a funny site and Surf Watch can verify that it's me and all that kind of thing, then sure, this would allow the window of vulnerability to be zero. Q Let me just, because this is such an important point, read you what you said last Friday in your deposition and whether you would still agree, it's on Page 165, Line 10: "I feel that there is some reliability problems in terms of using an exclusion list," that's with respect to parental control software, "keeping that exclusion list up to date is the biggest issue. Until the exclusion systems that I have seen are updated on a weekly or a monthly basis for their exclusion lists, and new sites are being generated all the time, and between the time when a new site is generated and the time the exclusion list update comes in there is a period of vulnerability," that's the period you're speaking to today, correct? A Yes. Q Okay. JUDGE SLOVITER: But the technical feasibility is there? THE WITNESS: Yes. JUDGE SLOVITER: And the only question is, I gather, the -- MR. BARON: The lag time. JUDGE SLOVITER: -- comprehensiveness of it? MR. BARON: That's correct, your Honor. JUDGE DALZELL: Well, and the desire. JUDGE BUCKWALTER: And the what? JUDGE DALZELL: And the desire. I mean, many of the plaintiffs in this case who some reasonable people might think are purveying, we'll use the motion picture parlance, NC-17 say we're not doing that at all, we're giving safe sex information, okay? Now, in the questions you're asking should they advise Mrs. Duvall we're NC-17 even though they don't think they should? MR. BARON: I am establishing through this -- JUDGE DALZELL: Is that what you're getting at? MR. BARON: Your Honor, that's a different legal issue and it is, I would submit, a legal issue. I am asking questions to the witness about a technical issue on the safe harbor provisions. THE COURT: Okay, fine. BY MR. BARON: Q Let's turn to another area of cyberspace and I regret that there's a whole new terminology associated with it -- JUDGE SLOVITER: Okay, you'll go slow. (Laughter.) BY MR. COGAN: Q UseNet, what is it? A I won't go into the history that I did when we talked last Friday, suffice it to say, it's a outgrowth of a distributed bulletin board system that started with computers calling each other up over the telephone and has migrated to providing the communication over the Internet. There's a few hundred thousand UseNet servers, they're just computers around the world running UseNet server software. They receive news group articles, which are just messages like E-mail messages, there's a characterization of a news group at the top of the article. News group articles are hierarchically organized, so it's -- one that I happen to read is rec.autos.sport.F1, because I happen to be a Formula One car racing fan, and so this is articles about car racing, about Formula One car racing, and it's put into a separate directory on the server. And then as I as a client can -- I as an observer can fire up a news client, which would then go off and I could tell it I want it to look at this subset, and it would show me the articles in the Formula One subsection. Q Okay, thank you. And what is known as NNTP? A Network News Transfer Protocol is the language which is used for the UseNet servers to talk to each other over the network itself, it's a handshaking mechanism by which a server tells another server I've got Article Number 1234 from Site 7, do you want it, and the other server can say yes or no. Q And what are ISP's? A Jumping around in technology a little bit here. ISP is the term that was actually coined by the National Science Foundation, it refers to Internet Service Provider, it's a company or an organization which is providing connectivity, Internet connectivity. It may or may not also include services such as news services or time services or E-mail forwarding or things that, but the fundamental service that it's offering is connectivity, the ability for Internet protocol packets to get from your local network to out into the Internet to -- theoretically to some other local network some place else. Q Are there approximately 15,000 global UseNet news groups? A There are somewhere -- there is actually probably considerably more than that news groups, as far as global news groups, it's a very hard number to determine because it depends on one's definition. I ran the news server at Harvard for a long time and I was getting Japanese news groups. Now, I would have stopped them except there were some people at Harvard who wanted to read the Japanese news groups, they were in transcribed Japanese, which I couldn't read at all, it looked like encrypted text to me. So, there -- it's hard to define. I would say that because I was getting those in Boston and they were being generated in Tokyo those are global news groups. If we're using that kind of definition there are at least that and probably more, but I don't know for sure. Q Would you say there are approximately 100,000 articles posted today? A That's a reasonable estimate for the ones which go out on the -- in that set of, quote, "global news groups." Q Now, this is a simple question, but how do you post an article on UseNet? A You compose a message, textural message usually on your UseNet client, which many of the browsers now include, and you say -- you tell that client which list of news groups you wish to post it to. The client then contacts the local server and says here is an article for news group rec.sport.autos.F1, and then hands it off to the server. Q Is there any difference with moderated news groups in terms of how an article is posted to UseNet? A There is no difference on how it's posted, what happens after it's posted is different. In an unmoderated news group when I do that posting to rec.autos -- rec.auto.sport.F1 my server would then automatically distribute it to all other servers which it had a communication with, which is at Harvard there may be a dozen different servers that it interacts with, so there would be about a dozen different computers it would send off this article to. And they would then propagate across the world, servers talking to their adjacent servers, just distributing it in an ad hoc interconnection mode, nobody controls that. In a moderated news group the posting would then go to my local server and then on that server it looks up and says, oh, this is a moderated news group, there is a list of moderators which is maintained on a few dozen sites which allow -- which would support the service of providing this forwarding list. My server doesn't maintain one now, I used to but I'm no longer in charge of that server, so I don't do this anymore. But the server that I would deal with would then look at it and say it's a moderated news group, I need to send it off to a server which contains a list of moderators, so it sends it off to one of these sites around the country -- around the world which contain the list. It would then go -- that site would then go through the list, forward this posting, which is really a textural message in my case, off to the moderator, which would then do whatever the moderator wanted to, including just automatically forward it into the news group or put it in their in box and read it, doing whatever the moderator wants to do and that would depend on the moderator, there is no set set of procedures or rules or software to support moderator functions. JUDGE SLOVITER: When I read your direct testimony I wondered about this, when you say a moderator you mean a two-legged, regular person? (Laughter.) THE WITNESS: Anybody -- JUDGE DALZELL: To wit, a human. JUDGE SLOVITER: Yes. THE WITNESS: Anybody who is willing to sit through, in the case of rec.autos.F1 it's now two or 300 messages a day, anybody who is willing to sit through two or 300 messages a day to decide whether they should be out I wouldn't necessarily call a regular person, but -- (Laughter.) JUDGE SLOVITER: I guess I asked for that. Who pays these people? THE WITNESS: These are -- almost all of these are voluntary efforts. There may be -- there are moderated news groups which are provided by corporations. For example, a company building some product may have a moderated news group which speaks -- talks about that product as a subset of the news group hierarchy which is specifically for business and it's a bus., dot, company name, dot, product, and they --those companies may pay a moderator to cut out redundancies or to answer the questions that show up in the mailing, whatever they want to do. But the vast majority of the moderators are volunteers. And there are quite a few moderated news groups, but the busy ones tend not to be because it's just too much of an effort for a volunteer to do. BY MR. BARON: Q But to summarize here, the moderator's role is to decide what messages are forwarded to the news group, correct? A That is correct. Q Could you describe the term hierarchy as it applies to the UseNet groups? A Hierarchy is just as the -- what I said, that the Formula One news group is in rec.autos.sport.F1. Rec. is a subsection of the news groups which are for recreation, autos is a subgroup of the recreational, which is dealing with autos; there's also sky diving and things like that in that same recreational. Within autos there's people who want sports, which is what I'm interested in, but there's also folks who do restorations of antique cars and there's a subgroup for them. And then within the sports category there's half a dozen or so different categories and the one that I happen to be interested in is Formula One. So, the hierarchy is that listing of -- it's the tree which winds up with a specific pointer to a specific news group. JUDGE SLOVITER: And that's a vehicle by which you get to precisely that which you're interested in? THE WITNESS: That is in theory the case. In practice people are a little less discriminate in what they post to news groups than perhaps they should be, but the aim is to make it so that the subgroup -- the news group is as closely focused on the topic you're interested in as possible. When I first started out doing this news group stuff it was -- rec.autos was the division and in rec.autos there may be two dozen messages a day. And then when that built up so that the volume was high they broke it up into -- under autos they put sport and restoration and et cetera to further subdivide it, in order to try and make it more and more focused. JUDGE SLOVITER: But the reliability is dependent upon whoever is labeling it -- I'm not sure that's the right word in -- JUDGE DALZELL: Posting it. JUDGE SLOVITER: Posting it, thank you. THE WITNESS: Posting. Whoever puts down on their browser, when they say post it they write down what news group it should group -- news group or news groups it should go into and it's entirely dependent on that person making the correct choice, that is correct. BY MR. BARON: Q So, let me just try to recap that. You said that some individuals might post indiscriminately to news groups that are sort of off-topic, but the point is that the individual poster controls where he or she will post the article to whatever the UseNet group is of the 15,000 -- A That's correct. Q -- and all of the hierarchies therein? We discussed the K-12 hierarchy in our -- last Friday, could you just tell the Court what a K-12 hierarchy is? A I know about the K-12 hierarchy only because they started -- they started it up at a time when I was running the Harvard news server, it's a sub-hierarchy that's specifically designed for people in the kindergarten through 12th grade with specific classes or specific topics. K-12, dot, one was for the first grade and they had topics relevant either to teachers or to students within first grade. Q Could you describe for the Court what the difference between the alt. hierarchy and the other hierarchy is? A Alt. hierarchy is the one which is, let's say, more --more traditional in the Internet sense of chaos. The other hierarchies, the rec. hierarchy, the science hierarchy, the -- there's a few dozen, K-12 hierarchy, et cetera, are hierarchies where there is an agreement amongst the people running the servers, on the main servers that they will have a controlled method for creating new news groups. And the controlled method is that somebody proposes a news group to a particular news group, which is about discussing proposing new news groups, and it's discussed on there and if there's enough support indicated by E- mail to the proposer that this particular news group should exist then the proposer can put in a news group creation request, which will then propagate across the net. One of the things that happens is there is a few places which maintain lists of, quote, "legitimate news groups" within different hierarchies and these lists are periodically posted to UseNet, to the UseNet as another article. The UseNet software can be configured to automatically review that list of legitimate news groups and delete any non- legitimate news groups, any news groups which do not appear in this list. JUDGE SLOVITER: What would be a non-legitimate news group then, just because it doesn't appear? JUDGE DALZELL: It's considered irrelevant? THE WITNESS: It's -- the structure is that, let's say, I wanted to create a news group on rec.auto.sport.F2, which is Formula Two. Well, there doesn't happen to be a lot of Formula Two activity these days. And after some discussion on the group -- on the new group list it was determined there isn't much support for that, and I go create it anyway, then the maintainer of the official list would say, well, that didn't get enough support, it didn't go through the right process to get that news group created, so that's an illegitimate news group, so I won't put it on the check list that goes out periodically. So then automatically when this check list goes out, some sites have set it up to automatically delete those unapproved news groups, others send mail to the news group operator or whatever. Alt. news groups do not have somebody who is maintaining that list of, quote, "legitimate," which means that news group are created ad hoc-ly by anybody, literally anybody in the hierarchy. So, there is a news group that's alt., dot, Swedish, dot, f, dot, borg, dot, borg, dot, borg, dot, borg, which I kept trying to remove, but it kept coming back. But there's nobody making any kind of check as to what -- any kind of list of what is a legitimate one. So, the alt. hierarchy is the old chaos of the Internet, free-will kind of hierarchy. MR. BARON: It might be helpful to look at an exhibit, if you would turn to Defendant's Exhibit 10 and see what we're talking about in terms of the alt. hierarchy. BY MR. BARON: Q I concentrate on the last two pages of this exhibit. The exhibit is from something called the Internet Yellow Pages, Second Edition; you've seen that book, haven't you, Mr. Bradner? A You showed me this same thing last week and I have seen earlier editions of this publication. Q Would it be fair to say that within the alt. hierarchy there's an alt. binaries sub-hierarchy? A That's one of many in the alt. hierarchy, yes. Q And there's an alt. sex sub-hierarchy A That's correct? Q Any particular ISP can decide whether to include the alt. sex hierarchy or the alt. binaries hierarchy, correct? A This was a question that you asked me last week and I maintained that you were using the term ISP incorrectly in this context. Any operator of a news server can determine what news groups that that news server will and will not maintain -- will and will not accept, and will and will not maintain. Some ISP's run news servers, some ISP's do not run news servers. So, to say that an ISP does this is an incorrect characterization, a news server operator can make that choice. Q Okay. Could you tell the Court what binary files are? A Binary is the computer jargon for a bit pattern which is used to represent any one of a number of things, for example an executable program, if you want a new helper AP, a new thing which draws pretty pictures on your screen when the screen is supposed to be idle, a screen saver, there are binary programs available to do that, you download them. In actuality UseNet only transmits printing characters, so in order to deal with the binary nature, the nature of non-printing characters, because the actual executables in the computer are stored in a eight-bit bit pattern which turns into gibberish when you try and print it, they actually translate each eight-bit character into two printing characters and then retranslate it back into -- you can retranslate it back into a printing -- into a binary pattern at your local site, on your local client. Q Just to be clear, can binary files include graphical image files, and I'm using that in the lower-case sense of the term? A Binary files can be, as you pointed out before when you were talking about URL's, they can include graphics files, motion pictures, sound, program -- pieces of program, sub-routines, but it can be -- you can get your voice mail via E-mail by including it in a binary file. Q Individuals can post binary files to any UseNet news group, correct? A In -- anybody can post -- Q Other than moderated groups, I don't mean it to be a trick question, I'm sorry. JUDGE SLOVITER: You mean you'll let him know when you do? MR. BARON: Right. (Laughter.) THE WITNESS: Well, I was going to catch you on that anyway. Anybody can post any file to any news group; if it is moderated, the moderator can control what goes in there. All files look the same, because they -- as long as they have the formatted point at the top, the formatted text at the top indicating a news group name and an article I.D., then they're in the correct format and the news servers know how to deal with them. The contents after that just look like printing characters, some of which are -- have sense to them and some of them don't, the ones that are binary tend not to. But so do the ones, for example, that are slightly distorted in order to make them not easily -- not trivially readable because it's, I don't know, a dirty joke or something, they have a very simple encrypting mechanism called Rot 13 or Rotate 13, it comes from one of the ciphers that Caesar used, as I recall. You just substitute its -- you take every letter in the alphabet and take the 13th one in a round trip -- or further along in the alphabet. And, so, that looks like gibberish too, but in actuality it's a one-character-per-one-character substitution. Q But it's an encryption scheme? A It's an encryption scheme. Q Let me -- forgive me if I'm being redundant, but you can also post graphical image files to any UseNet group, correct? A I think I just said that. Q Okay. Therefore, one makes a conscious choice when you post graphical image files or binary files whether you're going to post them to the alt. sex hierarchy, the alt. binaries hierarchy or any other place on UseNet, correct? A Just as one makes a choice when posting any article. Q Okay, thank you. May we turn to Defendant's Exhibit 12? (Pause.) Q Do you recall my showing you this exhibit on Friday? A Yes. Q Could you best characterize this, maybe you can do it better than I as to what this sort of artificial construct represents in terms of header information in UseNet? A It represents the basic UseNet header, which is present on all UseNet news messages -- articles, plus some things which are not in the basic. The ones labeled "mime version" and "content type" and "content transferring coding" and "X mailer" are ones which are not part of the basic set that's part of UseNet, it's the UseNet format itself. The others, the path is the sequence of computers that this article went through, and that path can be 30 or 40 or 50 computers long; the from is the stated name and E-mail address of the source of the message; the news groups is the list of news groups that the message was for; the date is the date; organization is the stated organization of the poster; the message I.D. is an important thing, because it is what is used to undifferentiate two messages which otherwise may look the same and make sure that messages don't loop around in the network forever, a data base is maintained of message I.D.'s which is relative to -- the message I.D. includes the source post's name, so Message 13 from this host is not repetitively posted to the news group accidentally. And the NNTP posting host is also not part of the original basic code, it's something that was added when NNTP came into play. Q Can we just hold that as a place holder here and explain for the Court what a news reader is? A A news reader is a piece of client software that -- in current environment most of them go off and speak NNTP to a news server, a UseNet server. Q Am I correct that some news readers are embedded in browsers? A Yes. Q And some news readers have the ability to do what you term threading, i.e. they follow articles based on the subject line of the posting, correct? A That's correct. Q Back to this exhibit, in theory an enhanced protocol for UseNet could include an extra line which essentially embeds content information, correct? A Yes. Q Thank you. Let me turn more quickly to other applications on the Internet. You have described IRC, could you explain for the Court what Internet Relay Chat is? A Internet Relay Chat is a way by which if I type on my keyboard it can appear on the screens of many people around the world simultaneously, and when they type on their keyboards it appears on my screen and other screens. Q I just have one question for you, Mr. Bradner: There are moderators or channel operators on IRC, correct? A In some cases there are, in other cases there are not. Q And those are human moderators, human channel operators, correct? A The only ones that I know of are. Q Okay. A Some I have question about, but... (Laughter.) Q All right, let's move to List Serves, could you explain briefly for the Court what they are? A List Serve is a -- actually a product name, it would be better to refer to it as an E-mail exploder. You send E-mail to a piece of software which then re-sends this piece of E-mail to a list of recipients, that list can be quite extensive. The ones I run on my local machine, I have an E-mail exploder for one of the IETF working groups, it has two or 300 -- maybe it's 180 now, I pruned it a little recently, different addresses that I have -- any message sent to that address, BMWG at Harvard, dot, EDU will be in turn forwarded to this list of addresses. List Serve is a particular product that implements this kind of E-mail exploder. It has some fancy features because it can deal with -- it can talk with other List Serves over the network and some con --regulation of what's -- which -- who -- which exploder has which addresses to forward to. But basically what you really mean is an E-mail exploder. JUDGE SLOVITER: So, once again, it's a vehicle by which one expands the recipients without the sender -- or the source necessarily knowing where it's going? THE WITNESS: Specifically that is the case. BY MR. BARON: Q I just have one question, Mr. Bradner: There are moderators on List Serves or E-mail exploders, correct? A I would say on the majority of them there are not. Q But there are some? A There are some. I actually currently, personally do not deal with any E-mail exploders that do happen to have moderators, all of the ones that I deal with are ones where I send mail to the exploder itself and it just forwards it. A moderator, I would send the mail to the moderator and then the moderator would in turn send it to the exploder list. I don't happen to deal with any, I know that some exist. All of the ones in the IETF, for example, for all of the working groups are unmoderated. Q Let's turn to E-mail and I just have one question: It's true, is it not, that some E-mail user agents allow you to separate out the mail based on the source of the message, the subject line of the message or a combination of those, correct? A That is correct. Q Have you heard of Eudora? A Yes, I have. JUDGE DALZELL: What's that? MR. BARON: Eudora. JUDGE DALZELL: As in wealthy? (Laughter.) BY MR. BARON: Q You have an extensive background in FTP, file transfer protocol, correct? A I'm not sure that it's an honor to say that, but, yes. Q You told me that last Friday. Conceptually, it is possible to block access to an FTP site, is it not, on an a priori basis by means of a password, correct? A Yes, if FTP is a way that I can sit at a client and ask to access to a server, an FTP server, and there are two ways to do that: One is what is called anonymous FTP, by which I give the log name anonymous when asked for my log name, my user name, and then I give my name as a password just to indicate for tracing purposes who I am, but of course that depends on who I say I am. This is the way that a huge percentage of the large data files, including for example the version of Flatland that I referred to earlier and all of the material on my machine are provided, they're provided by FTP. There is an alternate way, which is if I don't want to provide general access to some files then I can -- I can restrict that access to a password -- a user name and password protected, just as I restrict access to my local computer to people with -- that I have given accounts on the local computer to. Q The FTP protocol was standardized through the IETF RFC process, correct? A It was standardized very, very early, I wouldn't say that it was -- it was standardized early on, so I'm not sure that you could characterize it as going -- it definitely didn't go through the proposed and draft and full standard kind of process, it was one of the very first protocols on the Internet a long time ago. So, it way predates my involvement, so I couldn't speak with expertise on exactly how it was standardized, but my guess is some people got --sat down and said this is the way we're going to do it and, bingo, that's the way it was going to get done. Q Now, let me switch gears here. You told me last Friday that at Harvard there are many individuals who download to older versions of Netscape browsers for free, correct? A They download the version that is free. Q Okay. You stated last Friday in your deposition that the Internet is, quote, "becoming pervasive," unquote, do you stand by that statement? A And by pervasive I mean omnipresent, it is available anyplace. I can call from my hotel room, which I did this morning, and log in to read my E-mail. Soon I will be able to plug into a jack in the wall and identify myself and have Internet connectivity in the hotel room. So, it is avail --it will be -- it's becoming available wherever I want to go in and plug in and ask -- and identify myself -- connect to my home computer and then identify myself to the home computer with a log name password combination. And in that con -- the context in which I said pervasive I meant that it was becoming omnipresent, an ability for me to get it wherever I am. Q The Internet is also changing, correct? A Oh, at least. Q You recall that I asked you a visionary question last Friday, correct? A You asked me to make a speech and I did. (Laughter.) Q And I asked you to discuss with me where you saw the Internet going in the 21st Century, right? A Yes. Q And you expressed the view that there is not going to be an Internet as we know it today in the year 2000, correct? A The year 2000 or shortly thereafter, that's correct. Q Let me quote you from the deposition and ask whether you stand by this statement, you said, I'm quoting your speech, "Will there" -- it's Page 312, Line 19 -- Line 18, you said that you were giving a talk. "Will there be an Internet in the year 2000?" Line 19: "My conclusion is that in the year 2000 or shortly thereafter there will not be an Internet and by that I mean the Internet of today, that which people see and understand as the Internet is a differentiable data service. It's something that you see that is different than your television service, it's different than your telephone service, it's different than your fax service, it's a different thing than what you have, what you use for doing other functions. I believe that in the year 2000 or shortly thereafter we will have a unified general data service. In certain parts of the country already we have had a crossover between the amount of information carried as voice for the voice telephone network and the amount of information carried as data." And skipping down to Line 20: "This will become universal in the U.S. within the next half dozen years and there won't be something that you would say that's the Internet." Do you stand by those statements? A Yes. Q Thank you, Mr. Bradner. MR. BARON: I have no more questions. JUDGE DALZELL: But this unified general data service would act in similar ways that you have described both this morning and in your declaration, would it not? THE WITNESS: Yes. And what I meant by saying what I did was that right now you go and you go and buy telephone connection and telephone service from this vendor, and you go and you buy your cable service from that vendor, and you may go buy your electric utility from some other vendor, I predict in the future that you won't be able to differentiate between vendors, you will have a pipe into the house or maybe competition for pipes into the house and you plug this instrument onto it, onto this pipe and you get telephone and you plug this instrument onto the pipe and you get cable T.V. and you plug this instrument onto the pipe and you get whatever is the successor to the Web. And I believe there's a successor to the Web, I don't know what it is, but I believe that there will be some other way, some additional ways for a user to find things and interact with services around the globe and, in particular, doing that in a way which -- right now a great deal of the Internet is dependent on the voluntary efforts of individuals to provide material and I believe that in the long run that this -- the facilitating of this global, global and ubiquitous data service, one of the facilitating factors would be mechanisms for making it economically reasonable for content providers to provide content. That -- it's a real mixed bag though, I mean, one of the big things about a universal service like this is that it doesn't get controlled very easily. So, those environments where governments would like to control content, for example Singapore and China both have announced recently that they are working on figuring out ways to control content that their citizens can get over the net, over the Internet, the current Internet, this is a very big threat to that kind of their perception of what the social order should be. And I see this -- the Internet of the future being both a combination of a promise of tremendous reachability of availability of knowledge, availability of interaction, people interacting with people, and a threat to -- perceived threat to the ability to control what citizenry get, and that it is the balance between the perception of that threat and the reaction to the perception of that threat and the promise. I personally would rather focus in on the promise. MR. BARON: Thank you. (Discussion held off the record.) JUDGE SLOVITER: The Court thought that we would break now before you begin your redirect, to give you the opportunity to catch your breaths. MR. MORRIS: Your Honor, that would be fine. We have a somewhat unexpected scheduling problem, both the Government and we anticipated that Mr. Bradner's testimony would take a much shorter time than it has now. Mr. Bradner -- JUDGE SLOVITER: We didn't anticipate that. (Laughter.) MR. MORRIS: Mr. Bradner has a very important meeting relating to some international protocols negotiations, he -- in Washington, D.C. late this afternoon. He would absolutely be able to return first thing in the morning and, if it would be acceptable to the Court and the Government, we would suggest that we break for lunch and ask Mr. Bradner to return first thing in the morning. THE COURT: Is that congenital to the Government? MR. BARON: In theory, your Honor, it would be acceptable, but depending on the length of the questioning, it may be just for a few minutes and therefore it can be done. THE COURT: Well, what's your anticipation, Mr. Morris? MR. MORRIS: I think we probably would only go for 15 or 20 minutes. I don't know how many questions the Court might have... JUDGE SLOVITER: The Court thinks tomorrow morning? JUDGE DALZELL: That's fine. JUDGE SLOVITER: The Court thinks tomorrow morning. JUDGE DALZELL: I will have some questions. JUDGE SLOVITER: Is that all right with you? THE WITNESS: Yes. JUDGE DALZELL: Is that all right with you? THE WITNESS: Yes. JUDGE DALZELL: Okay? JUDGE SLOVITER: Even if you come back to lovely Philadelphia just for 15 minutes, you don't mind? (Laughter.) JUDGE DALZELL: It's on the way to Harvard. THE WITNESS: Yes. JUDGE SLOVITER: We will resume at 1:30. (Luncheon recess taken at 12:10 o'clock p.m.)
Return to the EPIC Censorship Lawsuit Page