(The following occurred in open court at 1:30 p.m.) COURTROOM DEPUTY: All rise, please. Court is now in session. Please be seated. MR. MORRIS: Good afternoon, your Honors. THE COURT: Good afternoon. (Discussion off the record.) JUDGE DALZELL: Going back to Dr. Bradner? MR. MORRIS: Good afternoon, your Honors. I'm John Morris, again, counsel for ALA plaintiffs -- JUDGE SLOVITER: And still. MR. MORRIS: And still. (Laughter.) JUDGE DALZELL: Dr. Bradner, you're still under oath. Do you understand that? THE WITNESS: Yes. JUDGE DALZELL: Very good. MR. MORRIS: Just to clarify the record, Mr. Bradner is not a Dr. Bradner, although -- JUDGE DALZELL: Okay, Mr. Bradner, all right. THE WITNESS: I may sound like one. JUDGE SLOVITER: You fooled us. JUDGE DALZELL: Sure did. SCOTT BRADNER, Previously Sworn, Resumed. REDIRECT EXAMINATION BY MR. MORRIS: Q Mr. Bradner, I really just want to go back over a very few points that you just talked about yesterday. Mr. Baron asked you about the Internet Protocol Version 6. Can you tell me, will the user of the Internet, typical user of the Internet, notice a significant change in how the Internet operates when Version 6, the next generation, is rolled out? A The IP Version 6 itself is just a replacement for the wheels, the bit -- the packet format of the packets that go across the net. The applications do not change in general by changing the underlayment which goes over. There is a possibility in the long term to get some increased functionality to do with the ability to do real time processing, telephone over the Internet kind of things, but this is research work and development work yet to be done. But basically the applications will remain the same applications you're currently using, and you wouldn't see any different -- the user would not see any difference. JUDGE DALZELL: So what is the change in some? THE WITNESS: The change in some is that the individual packet can address far more nodes, so therefore -- JUDGE DALZELL: Okay, so it's the 32 to the 128 bits. THE WITNESS: Yes. BY MR. MORRIS: Q And, Mr. Bradner, in your declaration submitted as direct testimony here, you describe a number of structural and operational issues raised by the Communications Decency Act. Would those issues still be in play and still be raised under Internet Protocol Version 6? A The issues here referred to are the inability of the speaker or the poster of some information to be able to determine whether the recipients are all qualified non-minors, is that the -- Q Right. A And no, they would not change. Q Okay, let me ask you, the inability of speakers that you described in your direct testimony, would that significantly change at all if you accepted Mr. Baron's proposal of placing a tag in an HTML file or a tag in a URL? Would the speaker's ability to ensure that minors didn't have access to speech significantly change? A I would not think it would make any difference to my level of comfort that my speech was so protected, because in order to agree that it was, I would have to hypothesize that all of the possibly affected minors were using web browsers which supported this tagging, and they all had been configured to have that configura -- the exclusion features turned on. And I would think that that would be a bit of a hard push to make the -- for me as a speaker to make that assumption, particularly since the pervasiveness from what I read of the statute which is tedious to read, though I've read it a few times, it doesn't constrain me to insure only within the United States. So I would have -- since all of my speech on news groups and things go worldwide, I would think that I would have to be assured that worldwide that every minor who happened to have -- might potentially have access would have such a browser and have it configured in a way which restricts access. And I think that would be a very difficult thing for me to assume, and I wouldn't want to assume that. JUDGE DALZELL: How about if you limited it to the United States? THE WITNESS: It's still something where the individual user, or the user's parent, if that be the case, would have to insure that each individual minor who has access through whatever means, through school, through libraries or through the home, would have a compliant browser and have that browser configured. And there's no way for me as a speaker to know whether that in fact is the case. And that's my -- the substance of my direct deposition was that, that I as a speaker have no mechanism by which I can go out and examine the capabilities and feature sets of what the listeners are using to listen to my speech. So I don't know whether they would have that restriction in. I could not know whether they had that restriction in. JUDGE SLOVITER: While we're on that point, I'd like to ask a question so that we don't have to come back to it. On page 33 of your declaration, you, in discussing tagging, you said it is impossible to imbed such flags or headers in many of the documents made available by anonymous FTP gopher and the Worldwide Web without rendering the files useless. JUDGE DALZELL: Which paragraph was that? JUDGE SLOVITER: Well, it's on page 33. JUDGE DALZELL: Oh, page 33. JUDGE SLOVITER: Yes, page 33. Is that part -- MR. MORRIS: It's paragraph 79. JUDGE SLOVITER: Oh. Is that part of the same matter that you're talking about? THE WITNESS: No. This is a somewhat different matter. The matter I was just speaking of is whether I as the speaker could tell whether a user was operating a browser that would inhibit -- that would watch those tags. What I'm speaking of in the deposition here and I also spoke of yesterday for a minute was that there are certain kinds of files, whether there be voice files or picture files or executables programs, computer programs, which you couldn't actually go in and modify by adding some character stream which would be these tags, without making them inoperative just simply because they are a particular format. And they expect to be in that format and if they're not in that format and the format would be destroyed by adding some tags, then they wouldn't work. That doesn't mean that a new format couldn't be derived. But the existing ones would not work. JUDGE SLOVITER: Technologically. THE WITNESS: Technologically, yes. BY MR. MORRIS: Q Yesterday, at least on our side of the table, a little bit of confusion arose in terms of the definition and the explanation of links on the Worldwide Web. Could you just take a moment and explain the relationship between a link, kind of a link -- the linkee and the linker. JUDGE SLOVITER: In simple language. (Laughter.) No, I really think that it is necessary if Judges are supposed to decide matters, then it is necessary for the parties and the witnesses to present it in intelligible form. THE WITNESS: Well, that's a strong challenge, but I will try. JUDGE SLOVITER: Well, but it's a necessary one. THE WITNESS: And if I fail, you will let me know, I am sure. JUDGE SLOVITER: Well, yes, and the people for whom you are testifying will know, so I think it's very important. THE WITNESS: Basically the way I would refer to a link in this context is simply a pointer. It is a piece of text, and it's just a character string, it happens to have some format to it, but it's a character string which specifies a particular computer on the Internet, and a particular file within that computer that is being pointed at. It also has some structure at the beginning which says what program do you use to go point at that. So I can say, use the Web or use FTP or use Telnet or use some other application-level program to go look at this file. So the file itself is -- although I used an example of one in my deposition of E2EVF12.IP which is one of the data files that's on my machine that's referring to the performance of some network device. That's a file name. The pointer to that would include the name of my machine, that's at NDTL.HARVARD.EDU and then the path to that file and then that file name itself. So all it is is it's a way to globally say this is the file that I'm interested in. It's not relevant to where the viewer sits. It is an absolute reference to the particular piece of data, and the particular piece of data is always a file of some kind, although with a reference of what program to use to do that. Now, I think that the question you're asking is --or the implication you're coming up with is do I have to know, if I'm the provider of a file, if I have a file on my computer, do I have to be involved with somebody pointing at that file, and the answer is no. I have no way of knowing whether somebody has a URL, one of these pointers, which refers to one of my files or refers to my computer. It can be -- other people do that. I know they do, I don't know who, but I do know that computer manu -- vendors of these devices have given out pointers to their data within my system. But I don't know who did it. I have no way of knowing. There's no involvement on my part. It is simply that somebody has said, if you're interested in finding out about the performance of this product, here is the URL of the file. And so there is no relation, there is no implied relationship between the pointer and the pointee, to use your semi-technical words. Was that clear enough? JUDGE SLOVITER: Yeah, finally. (Laughter.) BY MR. MORRIS: Q Just as I understand it then, someone else, a company can point to your computer and point to a file on your computer. Do they have to point to your home page, or can they point to other files on your computer? A As I mentioned yesterday, they can point to the specific file that they're interested in, bypassing any structure that I have, including my home page or any subsidiary pages or any directory structure. They specify the particular file that they want people to look at. It is irrelevant to what sequence I might want them to go through, if I had a home page and subsidiary pages, they would be bypassed entirely if the pointer so chose. If the pointers chose to specify a particular data file, it would bypass all of that. Q And so if you placed some HTML flag or tag in your home page, that would be bypassed by someone linking to one of your lower pages? A That is correct. This page, the home page, would not be loaded, would not be observed, would not be accessed, would no be hit, to use the parlance that was talked about this morning by the person who was accessing the data file. It just wouldn't be seen. Q And say -- in a lower page, say you placed a flag in a lower Web page, and that Web page included a reference to an image on your computer. You didn't place a flag on the image because you had a flag on the page that gave immediate access to the image. Could someone link to that image without linking to the page that you have intended to be the presentation of that image? A Yes. You can configure browsers such as Netscape such that when you are selecting an image or something like that, selecting a URL, it will show up in a screen, a little window in the page. So if you type that down, type down, write down that information, you have the literal pointer to a file. And if the file happens to be a picture file which is imbedded in -- which I would normally embed in one of my pages, then you can reference it directly. It is just another point to another file. There is nothing special about it in the context of the way things are organized on the computer, it's just another file that happens to have a pointer to it or more than one pointer to it, however many pointers that people, including me or others, have created. Q Okay, there's been a fair bit of discussion today and yesterday about the concept of a hit. In terms of you as a speaker being able to control access or screen access in some way to who's getting your files, does the -- if someone comes to your Web page and initiates ten hits by getting an HTML file and a graphic image and an audio file, to keep content from that person, would you need to screen just once, or would you need to screen each request for each different file? A Making the assumption which is implied in your question that there is a way that I can identify the originating query by using its IP address, and I want to somehow decide that this person or this computer where the query is originating, and that's what I'm using to filter on, then from the server's point of view I would have to make a check for that query every time a query -- a check for that address every time a query came in and a query is a hit. So for every hit, my server would have to verify that address. Q So if a thousand people on a given day visited your site and generated let's say 4,000 hits by looking around your site, then do I understand that you would have to -- one would have to screen 4,000 times or just 1,000 times? A The actual mechanism would require the screening 4,000 times. Now, just the software puts things in the temporary storage called cashing, such that the second look-up is quicker than the first one, but it still has to look up each one individually. Q Okay. Let me just run over a few little loose ends on some of the other areas that you were asked about yesterday. With listserves, can you just explain that there's an individual or perhaps a company that decides to create a mail exploding program, like Listserve or Major Domo. and then who actually speaks through that listserve? Is it just the company, just the individual who created the listserve, or who speaks to that? A Most listserves are for discussion. Some listserves are for propogation of information like advertising. In the latter case then the organization that creates it, it's just a mailing list for that organization, and in which case just somebody from that organization speaks. But for the vast majority of listserves or mail exploders, it is a communications mechanism between the people on that listserve. Anybody who sends mail to the listserve will --the mail will be forwarded to all members of the listserve, all addresses which are in the listserve. There are some listserves set up so that only people who have subscribed can send mail to it, and others that are set up so that anybody can send mail to it, but only people who have subscribed will see that mail. But there is in most cases, although than the moderated listserves where the mail will go to a moderator and then be forwarded, anybody who has the address of the listserve can send mail to it and it will be forwarded to all of the folks whose addresses are listed in the list. Q So it's a fairly open means of communication. A Yes. Q Let's briefly discuss news groups, Usenet news groups. There was actually a little uncertainty today, I believe you probably were in the courtroom, about number of moderated news groups. Do you have any information that you could provide the Court, any estimate of moderated news groups versus unmoderated news groups? A I can only estimate based on when I ran the news server at Harvard, which I stopped running about two years ago. The number of moderated news group was in the ten percent or maybe a little bit less range. My guess today, and this is mostly a guess, would be that it would be a significantly smaller amount. As I think I've talked yesterday, news groups and the number of people on the news groups are growing quite rapidly, so the requirement of effort of a moderator is also growing. So that to find people who are willing to wade through thousands of messages a day to decide to forward them is getting harder and harder. We don't have that many masochists, I guess. And therefore the number of moderated news groups is dropping. A number of news groups that used to be moderated are no longer moderated because they can't find somebody with the right degree of gumption to moderate it. Q And moderators are usually volunteers? A In most cases where it's not a news group that's set up by a particular organization to talk about their product, they are volunteers, yes. Q Okay. Yesterday there was also -- you were asked about the creation of new news groups. I believe you said the Alt hierarchy was fairly free, people could create things individually. But the other hierarchies, could you explain again who creates that? Is there one entity or organization that approves the creation of a new news group within a --say, for example, the Biz hierarchy? JUDGE SLOVITER: The what? MR. MORRIS: Biz, B-I-Z, as in business. THE WITNESS: One of the hierarchies I mentioned yesterday. Somebody within a news group, taking for example the one that I used yesterday, the REC.AUTO (ph) news group, the traffic got large on the REC.AUTO news group, and a number of people proposed to the news group that the news group be broken up into subcategories for people of different interests. A discussion ensued amongst the people who subscribe to the news group, and a proposal was worked out as a democratic back and forth process and a suggestion was sent off to a particular news group which is called NEWS.NEWS GROUPS or something like that, which is a specific news group set aside to discuss the creation of new news groups, and a discussion occurred on there. Then a call for a vote is given. And then whoever is doing the proposing for this news groups, which is just somebody of interest, somebody who's interested in it, collects up the E-mail votes that they get, and posts a list of all of the votes. It's this many votes in favor of this proposed news group and this many votes opposed, over some duration of time. If the votes are more than a certain amount, and it used to be that you had to have 100 votes in favor and no more than 25 against, -- I don't know what the numbers are today, that was what it was three years ago, and I suspect the numbers are much higher today -- then it's declared that that news group is a legitimate news group. And by declared, it's just the yes, it met -- everybody agrees it met the requirements so therefore it's legitimate, and whoever put in the original proposal would then issue the news group creation command and it would go off and create the propagate the creation of news group. Then the one other thing is that, as I mentioned yesterday, there are volunteers who maintain lists of approved, in the sense of this process approved news groups, within different portions of the hierarchy of the news hierarchy, like within the Biz hierarchy, and a volunteer will add it to their list. Then this list is periodically sent out and some news servers are configured to delete any news group in a hierarchy that doesn't appear on the list. So it's a democratic ad hoc process. BY MR. MORRIS: Q Okay. Briefly, could you just explain the process of registering a domain name with the Internet, at least in the United States, as I understand it. Has your domain name been registered with the Internet? The NEWDEV.HARVARD.EDU(ph)? A No. The domain name is divided up in different sections, as was explained this morning about the .EDU and .HARVARD and then NEWDEV is my particular machine. The portion of the domain name which is registered is the organization's domain name. So in this case it's HARVARD.EDU is what gets registered. Within HARVARD I have registered NEWDEV, but this doesn't get registered with the Internet, it is entirely within HARVARD. I have a domain name of my own for my little consulting company and I have registered that. But the computers within that I register with myself and I rarely complain so I get them registered quickly. (Laughter.) But that's only within my own domain. The domain in the simple case, it is something like HARVARD.EDU, it would be a little more complicated if we have a domain such as the .USTOPLEVEL domain which was also mentioned this morning, that two-level country codes are one of the types of top-level domains along with EDU and ORG. In the U.S. these are divided geographically. So for example the Web page that I mentioned yesterday of IETF -- WWW.IETF.ORD is actually on a computer that's run by a company in Western Virginia, and the company's name is CNRI, Corporation for National Research Initiatives. And their domain name is CNRI.RESTON.VA.US. And then there's computers within that. So it's IETF.CNRI.RESTON.VA.US. The CNRI.RESTON.VA.US is what was registered as the domain name but the machine within it is not. Q Okay, just one last question. Judge Buckwalter earlier asked Dr. Croneberger if he had all the money in the world and a clear definition of the patently offensive and indecent, would it be possible to develop a system. And I'd be interested if you could tell the Court, is that possible? Do you think that's possible? A As a theoretical exercise, I think it probably is possible. But for me as a speaker to be able to be insured that this was being met, the only thing I can think of, and I have thought about this a little bit, is we would have to have a mechanism by which all of the listeners, in every category, listeners to news groups, listeners to listserve, listeners on the Web, all of the listeners, all of the people retrieving information, would have to be able to provide some form of identification. Now, the identification wouldn't have to be more than a "Yes, I'm over 18" kind of identification, but in order to insure that that's reliable, you would have to have a mechanism by which you could create this identification such that it couldn't be easily forged. There is a concept that is being worked on that is called Digital Signatures, and the U.S. Post Office as an organization is looking into doing an identification check at post offices whereby someone would come in and identify themselves and they would get a bit pattern, a digital signature, which they could then use to identify themselves in commerce and in correspondence, perhaps even at some point legal correspondence. One could in theory add to that identification some flag in there to indicate whether they were over or under 18. But that would assume the mechanism were actually installed and in place of going -- of people going to the post office and getting these identities, and that you would have to have the identity before you could surf the Web. Without that kind of structure, without making the assumption of something in place so that I could be sure that a listserve could be created that would only accept subscriptions from those over 18, and I could query that listserve to be sure that it was one which was doing that, I think it would be hard. But in theory I could imagine a scenario by which this could be done, but it wouldn't take quite all the money in the world, as in the proposition, but I would have fun if it were a cost plus contract. (Laughter.) MR. MORRIS: We have no further questions. JUDGE SLOVITER: Thank you. Is there recross? MR. BARON: Yes, your Honor. JUDGE SLOVITER: Mr. Baron. MR. BARON: Good afternoon, your Honors. JUDGE SLOVITER: Good afternoon. RECROSS-EXAMINATION BY MR. BARON: Q Good afternoon, Mr. Bradner. Mr. Bradner, I'm going to give you a hypothetical. Let's assume you have a simple Worldwide Web Web site on a Web server with a home page. And the home page has a URL associated with it. And for purposes of this example, we'll just name it this generic name. The URL is HTTP://WWW.WEBSITE. Are you with me on that? A So far it's simple. Q Okay. Now, let's assume that the home page also has pointers to ten files on that Web site or Web server. And each of those files has a separate URL which is an extension of the home pages URL. A Yes. Q And let's just for the purposes of this hypothetical say that the URL for each of those ten files is HTTP://WWW.WEBSITE/FILE1 and then iteratively FILE1... through FILE10. Still with me? A Yes. Q Now, you have admitted in your prior testimony that it is a trivial matter for an owner of a Web site to imbed a tag or label in HTML source code in the header of that source code on the home page, correct? A On any particular page, whether it be the home page or not, as long as it's an HTML page that would be easy to augment it, yes. Q And as a technical matter, it would also be possible to imbed a unique tag or label in HTML source code for each of the files 1 through 10 with those separate URL extensions in my hypothetical, correct? A I'm a little confused here, because to me you may be mixing two things. One is the files themselves, and I just said that any file you could put -- any HTML file you could put such a tag in. And the URLs which are different things. You could put tags in the URLs also. Q I'm not speaking about the tags in the URL, but the tags in the files. So each of those files with a separate URL address would have HTML source code associated with that file and a tag could be put in for file 1, file 2, iteratively, file 10. A If indeed they were HTML files, yes. Q And the operation of that tag on any one of those files on that site would, assuming that it's a tag that's set up to be coordinated with parental blocking mechanisms, would operate to block that URL extension, that particular file on that Web site, that unique file, 1 through 10. A Assuming that it was A, as you speculate, set up to do so, and B, that the user was using a browser so configured to do that blocking, yes, all files are equivalent in that context. Q And any browser set up to look at the URL for a particular file 6 or file 8 or file 10 would be blocked if there was that mechanism for tagging and the browser was set to look? A Well, to be very specific, just to not be confusing here, what would happen is in your scenario that the browser would go off and it would retrieve the file, file 1, file 2, whatever, it would retrieve that file. It would actually go off and start to retrieve the file, perhaps even retrieving the whole file, then look at the HTML if it's an HTML file, then make the decision. So the fact that it has a URL is independent. It just happens that all files that are on Web servers have either real or implied or constructed URLs simply because as I just said, the URL is really just a way to identify a file. So we want to make sure that we're talking about two different cases, because yesterday we were talking about putting tags in URLs and we're talking about putting tags in the files, and I want to be sure that we've got those separated because they're different concepts. Q Okay. And in paragraph 30 -- sorry, paragraph 79 of your supplemental declaration at page 33, you mention the concept executable programs. Could you explain with some concrete examples what executable programs are? A The example I used yesterday was a screen saver, which would draw pretty pictures on your screen instead of burning your logo into the screen when the screen's been idle for a while. Many of those exist on the net, and you can retrieve them anonymously FTP or over the Web for free. They are there because the screen saver will come up and advertise some product or it's just because somebody thought that the picture they did was cute, or the function they did was cute. It was an intellectual exercise or a class exercise. So many of these are out. Hundreds of these screen savers exist and you can go retrieve them, but they are binary files. Q Do they usually have a .EXE suffix? A The .EXE is for a particular kind of binary file that's used on PCs. There are other extensions for other kinds of computers, for other types of binaries. Q Now, to the best of your knowledge, are any of the home pages of any of the plaintiffs in this lawsuit, do they comprise executable programs or files? A Since I don't know the list of the plaintiffs by heart or even by remembering having read it, the full list, it seems to be quite an extensive list, I couldn't answer, but it would seem to be unlikely that a home page would be a binary per se, though possible, because it could be a graphical image, but it wouldn't seem to be very likely. Q If one had an executable program in a file, is there anything about that mere fact that it's an executable program that would prevent the owner of the Web site from registering or otherwise making known the page where the executable program was to the many tens or hundreds or thousands of directories in cyberspace or to any of the companies that operate browsers or parental control mechanisms. Is there something special about executable programs? A An executable program when viewed by the Web is just another page. It happens that when the browser goes to try and retrieve it, it uses a different retrieval mechanism than it does when it's trying to retrieve text or HTML. But it's just another page, and one can -- and that page has a unique URL, and yes, one could register in theory any URL pointing to any file that was on any server, and there is nothing special in that context about a binary, no. MR. BARON: I have no further questions. JUDGE SLOVITER: Any more from the plaintiffs? MR. MORRIS: No, your Honor. JUDGE SLOVITER: Okay. Judge Dalzell. JUDGE DALZELL: Yes. I wonder in paragraphs 73 through 76 you talk a good deal about cashing, and I must say I really don't follow it. Would you just, using the standard that Chief Judge Sloviter put so well, could you just rehearse for us exactly what you mean by cashing? THE WITNESS: That standard was that it should be understandable? JUDGE DALZELL: You'll forgive the expression, yes. (Laughter.) THE WITNESS: Well, let's take the example I think I used. But the basic idea is you would like to minimize the number of times that you go off and you retrieve something, that you actually have to go to the actual server to retrieve something. Cashing is not a new concept. It's used in computers between the processor and disk so that if you want to reload the same file off of disk between the processor and memory if you want to reread some portion of memory into the processor, instead of going to the relatively slow memory or the relatively slow disk, it goes and gets it out of the cash. In the case of cashing Web servers, which is what I was speaking of in there, the same concept applies. Many universities, many organizations have as a function built into their firewall or built into a device sitting between their network and the rest of the Internet, they have an ability to watch the HTML requests go by and see what comes back and save a temporary copy of what comes back. JUDGE DALZELL: How temporary? THE WITNESS: That's usually locally configurable and it's frequently done based on how much disk space you have. Once you start getting low on disk space, you throw away the things which haven't been accessed for the longest period of time. And -- JUDGE DALZELL: Well, you give in paragraph 73 the hypothetical of cashing when a user in Europe requests access to a Worldwide Web page located in the United States? THE WITNESS: Yes. MR. MORRIS: Your Honor, could I -- JUDGE DALZELL: Yes, sure, sure. THE WITNESS: Yes. JUDGE DALZELL: Now, do I understand correctly that the first time that the European user accesses the American information page, it is put somewhere in cyberspace on the east side of the Atlantic? THE WITNESS: In this particular hypothetical, to use a word that was just being used at me, what would happen is that the people who operated the transatlantic link, and it happens that in the real world of data networking the Europeans have to pay for their links into the U.S. The U.S. doesn't pay for the links because everybody thinks we're more important and so far the Europeans have agreed with that. I think that may change in time, but they want to minimize the utilization of that link. They don't want to have it retrieve the same file again and again and again from some U.S. site. So they would put on their end of the link a cashing Web server. So that when a request came in from Europe, if it wasn't in the cash it would go do the request across the Atlantic, across the relatively small and very expensive link, retrieve that file and then store it in the cashing Web server and pass a copy onto the requester. JUDGE DALZELL: But storing it in the Web server in your hypothetical in Europe? THE WITNESS: That's correct. JUDGE DALZELL: Now let's reverse the hypothetical because we're told a significant percentage of the sexually explicit material comes from abroad. I mean I know you're not an expert on that, but that's what we're told. THE WITNESS: I don't know from personal experience, no. JUDGE DALZELL: Right, but that's what we're told. Let's assume that's true, okay. THE WITNESS: Mm-hmm. JUDGE DALZELL: So the American server, the American user types in "sexy European girls," okay. Now, that goes to Europe, pulls it back and it's held for a time in somewhere on this side at the west side of the Atlantic so that it could be accessed again for some finite period of time? THE WITNESS: The European, in this case because the transatlantic link as operated -- as paid for by a European company, if the European company put such a server on this side in order to preserve the band width coming the other direction, which they could easily do, then yes, it would be stored for a time on this side and it would be identified by the URL, that's basically -- URL being index into it. JUDGE DALZELL: Same hypothetical except the provider of the sexually explicit material is in Thailand. Any difference in your answer? THE WITNESS: It's probably more likely to have a Web cashing server there because the transpacific links are very, very expensive, million dollars a month for a good-sized link from the U.S. to Japan, for example. JUDGE DALZELL: And who is paying for that link? THE WITNESS: The Japanese Internet service providers. JUDGE DALZELL: Collectively. THE WITNESS: There are a number of different links. I was quoted the million dollars a month quote by somebody involved in the Asian-Pacific Network Information Center a couple of months ago, and I'm not sure if that was from a particular Internet provider in Japan, or whether that was some conglomeration. JUDGE DALZELL: Judge Sloviter wants to ask about cashing. JUDGE SLOVITER: Now, given now that you have explained cashing in terms that maybe we can understand, thanks to Judge Dalzell and to your willingness to do so, what happens, what is the effect? Now that we know the concept, what is its relevance to the issues before us in this case? THE WITNESS: The primary relevance, the reason that I brought it up in the statement is that once this is cashed, you know when someone asks to see it, the original server does not see the request. The original server's request --the user's request is fulfilled by the cashing server and that request is not passed on to the originating server, so the originating server has no way of knowing who or how many or anything else. We were talking about hits earlier. The originating server would have no way of knowing how many times this file was retrieved because it wouldn't get the hits. JUDGE SLOVITER: And the file would be identified though as coming from the original server? THE WITNESS: It would -- remember that in the -- we were talking -- I missed yesterday afternoon, but I was told about your adventures in cyberspace -- JUDGE SLOVITER: And never get it back. THE WITNESS: Everything is just a URL. There is nothing particularly in that URL that tells you really where it came from. There's some names in it which if given some work we could figure out where some particular site exists, but we don't know where SEXYEUROPEANGIRLS.COM might be, the server might be. That actually might be in the U.S., it might be in Europe, could be anywhere. There's just the URL, which is, as I explained a minute ago, it's the application followed by a site name followed by a file name. What is saved in the server, in the cashing server, is the URL and the file associated with it. So that's all that's there. We don't have any way of tying anything else together than that. JUDGE SLOVITER: So you don't -- you can't identify either the pointer or the pointee, is that the point? THE WITNESS: It's not so much that, it's that if somebody in Europe were retrieving a data file from my server, let's say one of my newspaper columns, and they retrieved that, I would see a retrieval from the local end of the -- the cashing server in Europe or in the U.S., however it was done, I would see one retrieval from that. There could be many other people who might want to read it --that's not borne out by what my editor says -- but there might be other people. I would never know whether anybody else read it, retrieved it within the time period of how long the cashing server kept it. JUDGE SLOVITER: And would the cashing server know how many people retrieved it? THE WITNESS: As long as it in turn wasn't behind another cashing server. Because organizations put cashing servers on their boundaries to minimize the link, the utilization of their link to the rest of the Internet because those links are expensive, and if they're keeping, retrieving the same files, they're making use of the link that they wouldn't otherwise need to. JUDGE DALZELL: You see, this is why I think this may be important to our consideration. Whoever it is who created the Web page, let's say SEXYEUROPEANGIRLS, let's hypothesize in Luxembourg, does not think or may not be thinking about complying with Title V of the Telecommunications Act of 1996, but the cash server in the United States, as I understand your testimony, effectually domesticates that Luxembourg Web page, does it not? THE WITNESS: If indeed there is such a server there, but again, it could be a server there, it could be a server that's on the boundary of a corporation or a school. There could be many servers up and down the line. Each one of those, if they're within the U.S. could yes, be domesticating it. JUDGE DALZELL: Could Mr. Coppolino and his troops find that domestic cash server to prosecute them? And if so, how would they do it? THE WITNESS: You can -- actually there's nothing in the information that you retrieve when you -- when somebody goes to retrieve it, there's nothing to indicate where that came from. So if indeed for example the servers have been tweaked so that they didn't save copies of files from SEXYEUROPEANGIRLS.COM, then you wouldn't as someone reading from this side of the Atlantic, you wouldn't be able to tell that it was or was not cashed anywhere along the line. JUDGE DALZELL: And there would be no way for me to know that and therefore no way for Mr. Coppolino and his troops to know that? THE WITNESS: That's correct. JUDGE DALZELL: Okay. But nonetheless I would see an untagged SEXYEUROPEANGIRLS; correct? THE WITNESS: You could see that. JUDGE DALZELL: Even if all domestic information providers agreed, yes, let's tag all our URLs, the folks in Luxembourg don't give a darn what our law says, but nonetheless their information may come to my son, who's ten. THE WITNESS: That's correct. JUDGE DALZELL: Okay. THE WITNESS: The other worry that I have in that actually is a little bit perhaps even the other way. From what I understand, we went through this process of the CD providers agreeing to tag rap songs with parental guidance. And the experience was that those CD's so tagged sold better than the ones that were not. (Laughter.) I would have some worry that people would actually, SEXYEUROPEANGIRLS.COM would actually be quite happy to put tags on and then make sure that there were browsers readily available that would seek out sites that had such tags. JUDGE DALZELL: Okay, that's very helpful. A couple technical questions. FTP and gopher. Are they forms of search engines? THE WITNESS: No, they're just file retrieval devices. JUDGE DALZELL: And how do they differ from a search engine like Alta Vista (ph)? THE WITNESS: Alta Vista builds up a database which when you send a query to it, it looks into its local database to see whether it has entries that match that query, and return to you the pointers, URLs, that correspond to those matches. JUDGE DALZELL: As I understand, FTP and gopher or Alta Vista, if I type in Scott Bradner, they're all going to go looking for anything mentioning you, won't they? THE WITNESS: FTP -- no, FTP is an interactive program with which you can tell it to retrieve a particular file. So if I gave you the name of a file on my machine, there is a process by which you could interactively go through and retrieve that one file. It is not the kind of thing -- FTP is not the kind of thing, nor is gopher the kind of thing which by itself you can say, go find every file which has Scott Bradner in it. All it does is say, here is a file name. Here is a URL. Here is the extended URL. In the case of FTP, you actually give it the machine name separately and the file name separately, and you have to do all kinds of -- JUDGE DALZELL: But in Alta Vista I could just search Scott Bradner? THE WITNESS: It has been done. (Laughter.) JUDGE DALZELL: I'm sure of that, maybe by people in this room. (Laughter.) Lastly, you talk in paragraphs 70 and 71 of your declaration about common gateway interface or script, a CGI script? THE WITNESS: Yes. JUDGE DALZELL: And that you say is an option that is at least possible on the Worldwide Web where the -- if you had the resources you could screen out users? THE WITNESS: It's basically -- and we were talking this morning, we were talking about forms type interface where you have little boxes to fill in, one of the things that can happen when you fill in those little boxes is you can fire up, you can start up a program and CGI is one of the ways you can start up a program, which takes as its input information in the boxes. So for example it could take as input somebody's credit card number. JUDGE DALZELL: All right. THE WITNESS: And then if mechanisms existed to verify that credit card number purely electronically or over the Net, it could go ahead and do that in theory. JUDGE DALZELL: Okay, that's very good. Thank you very much. JUDGE SLOVITER: I'll ask a question while you're thinking. JUDGE BUCKWALTER: No, you go ahead. JUDGE SLOVITER: When you used the term "trivial" yesterday a number of times, and I'd have to go back and retrieve and I don't have it all here and we don't have daily transcripts so I don't know, do you mean by trivial, trivial to somebody with your expertise, or technologically trivial, or do you mean trivial because it doesn't take a lot of money or because it doesn't -- I didn't quite understand what you meant, and you used it at different times, so you may have used it differently. THE WITNESS: I believe most of the time that I used it it was in response to a direct question from Mr. Baron who used it in -- when we spoke last week. But be that as it may, I did use it and by that I really meant that it was really easy to do, given -- I think in the deposition what I said was -- or in the discussion we had last week, I said something along the line of given a sample home page with a fill in the blank of how you would designate naughty stuff below, anybody could take that sample and modify it for their own environment just trivially, because all it is is an editing task. It's just like editing a letter to one's mother. It's just characters. And as long as you have a model to follow and a set of instructions to follow, it's easy to do. So I think in most of the contexts that I was referring to yesterday, in that it's easy to do, not necessarily for somebody with a great deal of expertise but given the right environment, it would be easy for anybody to do. It would be difficult for me to do it today because I don't happen to be an HTML expert. I went out -- after our discussion the other day I went out and bought a book because he asked me some embarrassing questions I should have known the answer to. But it would take me some time and it would be untrivial, both in terms of time and at my normal billing rate expense for me to set up a home page to do this. But if I were given a sample home page with, if you fill in your name here, everything will be fine, it would be truly trivial to do. JUDGE SLOVITER: Thank you. JUDGE BUCKWALTER: I was thinking of carrying Judge Sloviter's example being simple maybe to the ridiculous, I hope not, but maybe my analogy would help me, when I was a young man I wanted to be a user of the local bar when I was in college. But the provider (laughter) was by law forbidden to serve me if I were under the age of, I think it was 21 back then. Do I understand that it's somewhat analogous that if the provider puts up a sign there that says no one under 21 allowed to drink that that's a tag? THE WITNESS: That would be -- that would be an equivalent to a tag, yes. JUDGE BUCKWALTER: That would be equivalent to a tag. And do I further understand that if that sign were in Japanese, I wouldn't be configured to understand that tag or -- is that what we mean by being configured? Make the assumption that I don't understand Japanese language, but -- THE WITNESS: Well, I think that's a common assumption in this room. JUDGE BUCKWALTER: Yes. THE WITNESS: Certainly an assumption that falls with me. Actually what I think I was referring to was a little bit different than that -- JUDGE BUCKWALTER: Okay. THE WITNESS: -- in that it was more along the line of yes, I guess, am I configured to be able to read and understand that sign. JUDGE BUCKWALTER: Me, the user, I'm not configured, so on -- THE WITNESS: Yeah, can you read the sign. JUDGE BUCKWALTER: -- that point, that tab would be no good to me. THE WITNESS: That's correct. JUDGE BUCKWALTER: Right now, I'm configured to understand English so the tab would be good. THE WITNESS: Yes. That's what standards are useful for. JUDGE BUCKWALTER: However, it wouldn't be sufficient to guarantee that I wouldn't get in the bar and get a drink, right? THE WITNESS: In the context of the law, I think what we're to extend this simile to probably ridiculous levels, it would appear to me that the provider of liquor to that bar has to insure that the bar doesn't serve underage people. And that's the paradigm. It would seem to follow from the law that it's the provider of the materials that has to insure, -- JUDGE BUCKWALTER: Right. THE WITNESS: And the liquor provider of the bar can't be sure that they're checking IDs at the door and -- JUDGE BUCKWALTER: He can have a gateway though which is the bouncer at the door who checks the ID or the password of the -- THE WITNESS: But would the liquor provider have to have that bouncer, or would the bar's owner have to -- JUDGE BUCKWALTER: Well, under the example I'm giving and the way it worked back then, the liquor provider had to have the bouncer. And isn't that what we're talking about here under this law, that the Government is saying the provider has to have the bouncer? THE WITNESS: The liquor, the Jim Beam has to have the bouncer, not the bar has to have the bouncer, and that's the difference -- that's my extension of your scenario to this law. JUDGE BUCKWALTER: Okay. Now, extending my scenario further, are you saying that that's very difficult to -- THE WITNESS: It's very difficult for Jim Beam to know that the bouncer is there and to know the bouncer is following the rules set which will result in underage people not being allowed into the bar. JUDGE BUCKWALTER: Okay. I don't want to go any further with this, because it takes up too much time and might not be particularly helpful, but thank you. JUDGE SLOVITER: I have one more basic question for him. While you were willing to do some explaining, and if we've done this twice, let me know, but I was trying to understand the use of the password system rather than the anonymous system. We haven't gone over that today, have we? THE WITNESS: No, we haven't. JUDGE SLOVITER: In different language -- which you talked about in your declaration. THE WITNESS: Yes. JUDGE SLOVITER: And I think you say that use of the password system rather than the anonymous system automatically would give the user access to the entire system. THE WITNESS: I do say that, yes. JUDGE SLOVITER: Yes. And if you would just quickly explain that, if possible, and then tell us why couldn't the access be restricted also? THE WITNESS: What I was explaining there was the existing software. Certainly software can be changed to do anything that you want it to do, given sufficient money and motivation. But the existing software for FTP which is what I was specifically referring to, anonymous FTP when you log in as anonymous, which is if you wanted to retrieve a file from my system and I gave you the name of the file, you could just go in using FTP, log in as anonymous, and then go to the directory the file is in and retrieve it. But when you do that, the FTP server, the demon, the server itself automatically restricts you to a subset of the information that's available on my machine. It's a subportion of the directory tree. So you can only see that, and you cannot change directory out of that. There's just --the software prevents you from doing that. It's called rehoming. It gives you a new home that's very, very restricted. In the currently available software, the FTP demons, that if you go and you log in as a regular user, which is what you do when you're logging in with a user name and a password, a legitimate one, my own user name and my own password, for example, if I were to do that, then it gives me access to the machine as if I had logged into the machine as a terminal. It gives me the same view of the machine. And that view of the machine is all of the files on the machine with whatever protections those files have for every user. So that if a file is publicly readable for all users, then it would be publicly readable for the FTP who was coming in on a log name and password basis. If it's not publicly readable, then it wouldn't be. What I was referring there was that it lays open for potential abuse the entire structure of the server site. In my case -- JUDGE SLOVITER: Of the server. What was -- THE WITNESS: The server computer, the server site. JUDGE SLOVITER: And you concluded from that in terms relevant to this case, because I'm not sure which side that goes for, but what do you conclude from this? THE WITNESS: There were two messages I was trying to convey there. One was that it would be an administrative burden for me to maintain a log name and password file for all of the very many people, individual users, who would want to come and look at and retrieve information from my site. So it's just purely an administrative nightmare to do that. And the other is that it would pose to me a security risk to make all of these files open, because I know that in my machine as in almost all machines I have not been perfect in setting the machine up, and I may have made some errors, and some files which should have been protected are not protected, or there is bugs in some of the software. And people could use the visibility they have into my machine to circumvent the security on it. JUDGE SLOVITER: And you're not saying that software couldn't be created that would do that, would take care of this problem, -- THE WITNESS: Absolutely not saying that. JUDGE SLOVITER: -- but it just hasn't been right now. THE WITNESS: It is not -- the current software does not do this. Future software wouldn't change the administrative burden tremendously, but -- JUDGE SLOVITER: Well, unless future it may do that, too. THE WITNESS: The future is unseeable. JUDGE SLOVITER: Okay, thank you. Mr. Baron? BY MR. BARON: Q I have just a couple questions. The record may be clear on this point, Mr. Bradner, but we were talking about the trivialness or the ease by which a tag can be put in HTML source code. And do I understand it from your testimony that those individuals who know HTML source code who would be designing Web pages, for that class of individuals it would be an easy, straightforward trivial task to imbed a tag in a header? A Actually I was going further than that and saying that if you gave me a sample of what it's supposed to look like, even if I don't know HTML, it would still be trivial. Q Okay. And this is a very important point that your Honors have raised about conventions and standards for looking at tags and headers. And so my question is that you recall your testimony yesterday about Netscaping approximately 80 percent of the browser market, correct? A As I recall my testimony yesterday was that Netscape claimed to be 80 percent of the market. Q I stand corrected. If -- and we spent a lot of time yesterday at the beginning of your testimony talking about the IETF and the fact that the Internet is a standards-rich environment, do you recall that? A That's a standards development group rich environment. Q Okay, all right. (Laughter.) A Which is slightly different. Q There are lots and lots of standards. A Yes, some of them conflicting. Q And draft standards, proposed standards, we covered all that. A Consortium standards, the end of the month club. (Laughter.) Q Now, if a Government agency or alternatively if in a consortium of private parties got together and adopted a convention or a standard for tagging in HTML code, that would essentially operate, like the movie rating system, be an adult tag, that is a technical matter, the browser market as just is a matter of technical feasibility the browser market represented by Netscape and Microsoft and others, could certainly pick up that tag that would be adopted as a standard or a convention, isn't that correct? A Yes, and they could even program the browsers to accept more than one standard for the same sort of information. MR. BARON: I have no further questions. JUDGE SLOVITER: Any further questions? MR. MORRIS: No, your Honor. JUDGE SLOVITER: Okay. JUDGE DALZELL: But I want to follow up on that. You said and Dr. Hoffman said that the Worldwide Web was not created here, it came out of Czern (ph), right? THE WITNESS: That's correct. JUDGE DALZELL: And so what Mr. Baron just asked you about hypothesizes that there is a plenary group that sets standards and by setting standards, doesn't that then exclude the possibility of new technologies such as the Worldwide Web which arose spontaneously, not even in these shores? THE WITNESS: That is a very insightful question. (Laughter.) That happens to be one of the subjects of the meeting I was at yesterday and couldn't be here on. The tension between standards bodies is a very serious issue. I made light of it just a second ago, but it is a very serious issue and there are many places in the globe where the development of standards and the control of the standards development process is seen as a strategic necessity on the part of some governments. And so the defining of what standards to use, how to develop them and what -- the level the mandating of those standards is seen as strategic. And it is absolutely true that a too-strong environment saying all standards must come from standards group number two has a serious impact on innovation. The Web rose out of a hole. The whole concept of the Web rose out of a need that we didn't know we had. We didn't know we had this lack of ability to do easy browsing because we didn't have the concept of easy browsing. This was something that sprung out of innovations on some parts of some individuals. It was not part of a standards effort. It was people doing something. It caught on because people wanted the -- saw it as useful. There are other holes in the Net. We don't know what they are. There are other needs that we don't know we have. And certainly being too reliant on "we only do those things which are standard" will stifle that innovation and it would be very bad for us. We went through many years of telecom where we did not have, let's say, rapid innovation because of that kind of centralized standards development constraint. So in the context though of what Mr. Baron was asking, if a consortium were to develop a tagging structure, it wouldn't have to be the same consortium that would develop future extensions on HTML. I met with somebody on Tuesday where the IETF is going to actually have a document which is the extensions on HTML and this is being put together with the W3 consortium. It's a nice conglomerate effort. But a group putting together a standard for labeling within HTML or within URLs or within copies of any text that happens to be around wouldn't necessarily have to cooperate with any other standards body in doing so. It could just create it. JUDGE DALZELL: And indeed, isn't the whole point that the very exponential growth and utility of the Internet occurred precisely because governments kept their hands out of this and didn't set standards that everybody had to follow? THE WITNESS: Well, it's actually even a little bit more contorted than that because the governments tried to. The U.S. Government and many other governments attempted to mandate a particular kind of protocol to be used on worldwide data networks, and this is the OSI protocol suite. The U.S. Government mandated its use within the U.S. Government and with purchasing material with U.S. funds. This was mandated in many European countries and in Canada and many other places around the world. That particular suite of protocols has failed to achieve market success. What achieved success was the very chaos that the Internet is. The strength of the Internet is that chaos. It's the ability to have the forum to innovate. And certainly a strong standards environment fights hard against innovation. JUDGE DALZELL: Thank you. JUDGE SLOVITER: Thank you. I'm sure counsel would like to continue the dialogue, but are you content to let it be at the moment? MR. MORRIS: We are, your Honor. MR. BARON: Yes, your Honor. JUDGE SLOVITER: I guess you are. I'm not (laughter)... We'll close now to resume on the 1st of April -- JUDGE DALZELL: On the 1st at 9:30. JUDGE SLOVITER: Although counsel is I gather going to meet with Judge Dalzell. Do they know that -- JUDGE DALZELL: Could we do that at 3:00 o'clock in the robing room, if that's convenient? JUDGE SLOVITER: Okay? All right. Thank you all very much, and have a good week until then. THE CLERK: All rise. (Proceedings concluded at 2:40 o'clock p.m.)
Return to the EPIC Censorship Lawsuit Page