Joyce Reynolds (jkrey@isi.edu) Interview -- short by Dennis G. Allard (allard@oceanpark.com) February 16, 1996 Information Sciences Institute Marina del Rey, California Appeared in OnTheInternet, September 1996 OTI: How did you get involved with work on the internet? Reynolds: I was looking around for jobs and I seemed to have a proficiency with computers that Steve Crocker here at ISI then saw. So I started out on an entry level. OTI: You were still a student at that time? Reynolds: I was a graduate student in history. I took background classes in programming to catch up with the computers. My parents both have science backgrounds but they're in the dentistry and medical fields. I seemed to have enough of a science proficiency that he said that this could work. This was back in the ARPA days. It was in July of '79. To jump ahead, Jon Postel took me on in March of 1983 to the projects we've been working on together since. OTI: What had you been doing up until then? Reynolds: Mostly proposals. Not on a Ph.D. level, but I could at least push papers and look at proposals and I understood these things. So, with Jon we started doing the assigned numbers documentation. OTI: About what year was that? Reynolds: August of '83. It was not called the IANA until 1990. Basically, the point of contact would be to send a message to Joyce Reynolds or Jon Postel at the time. Meanwhile, he had a need to update the File Transfer Protocol. So I helped him to update FTP and the Telnet specifications. OTI: So way back in '83 you were laying the foundations for IANA (http://www.isi.edu/div7/iana/), much in its present form? Reynolds: The evolution was that he took me on because how busy it got to be. Jon is the IANA. I noticed that Jon was really getting backed up with RFC's, Request For Comments, (http://ds.internic.net/ds/dspg1intdoc.html) because he had been the RFC editor since its inception in 1969. So I went down there one day to his office and I said you know you're going to have to learn to let go. The next day he comes wheeling into my office on a big terminal stand, piles of folders. They were all RFC's that needed to be processed, read, tagged, gone through, edited, published and he wheeled the whole thing into my office and smiled and said, 'I'm letting go', and walked away. So in October 1987, I started working with him and he just had this smile and I'll never forget this I stood there and I went "Oh, shit." And that's when I started helping him. What had also happened by this time was that the internet protocol number assignments had gotten too administrative for us and we had shipped all the management of assigning network numbers and system numbers up to what was then called the SRI NIC in 1986. That left me with enough time to take on the RFC's. OTI: Tell me how the SRI NIC got started, what would be called the InterNIC today. Reynolds: The SRI NIC had been in existence for quite a number of years. We delegated IP network numbers and autonomous system number assignments to the InterNIC and to other global repositories but the buck stops with Jon and me. OTI: The RFC approval process involves a lot of filtering up through the various IETF working groups. How does that process work? Reynolds: I'm in on that because I'm a member of the Internet Engineering Steering Group (IESG) of the IETF (http://www.ietf.cnri.reston.va.us/). I'm the area director for User Services. The RFC's are a publication vehicle for the standards of the internet community. For example, PPP is in what we call a standards track. And the IETF is the community where these things are developed. The IESG is the steering group for the IETF. Documents that are being developed in the IETF are called internet drafts. The IESG nurses the internet drafts. When they arrive here at the RFC editor's office they are still internet drafts. They don't become RFC's until we put a number on it and we announce it to the world as an RFC. OTI: How is that decision made? Reynolds: Let's say I have a working group in the IETF working on an internet draft. My working group chairs come to me and say this document is ready to submit to the IESG for approval. The IESG as a body gets together and teleconferences twice a month or more often. As the area director, I tell the IESG secretariat that I've got a document to send up to be an RFC. A notice is sent to the other area directors asking that they read the document. We'll have a teleconference. Usually things go through very well, there's other times when there are technical problems. A ballot is then sent out. Once we get through bantering and the ballot passes in the IESG, a message is sent by the secretariat to the IETF community and to the RFC editor pointing to this particular internet draft document saying this has been approved by the IESG as a Proposed Standard. Jon and I grab the internet draft and read it with an RFC editor hat on. We don't just do copy editing. We read it for technical soundness. Usually there isn't a problem but every once in a while Jon will find something he doesn't like technically and he'll state a concern. And that's inherent in him, he's just such a guru at things. But most of the time when it gets past the IESG, and accepted by the internet communities, it's the RFC editor's job to officially get it out as an RFC (Request For Comments) document. OTI: I read recently in a popular internet magazine that even when something becomes an RFC, that doesn't make it an internet standard. It becomes an internet standard when people start using it all over the net. So, what's the evolution of an RFC? Reynolds: We have three levels: Proposed Standard, Draft Standard, and Standard. If your document gets approved and is issued as a Proposed Standard RFC, it's off and running for a period of four to six months, sometimes a year. The vendors who are interested in this protocol immediately start testing and debugging it at their sites and they'll interact with other sites. Once everybody agrees that it is a relatively stable protocol, it goes to Draft Standard. The way you go to Draft Standard is you go back through this process again. There is a NEW document showing the testing and implementation, because everything is always a test. You've got to test and re-test. The same process will go back through the IESG, voting, etc. A new document which will obsolete the Proposed Standard document lands on our desk and the secretariat of the IETF will send a message to the RFC editor saying the document has been approved as a Draft Standard. OTI: The Draft Standard has its own set of tests? Reynolds: By the time it gets to Draft Standard, it's real, it's arrived. You never know what else you're going to find. Sometimes things never make it to Standard. The IETF is a volunteer group. There are times where it's either job changes or 'I don't have funding', or 'my employer won't let me do this any more'. PPP is one of the primary ones that I saw a real push where even if people fell by the wayside there was someone else that was going to keep going. There was some back-tracking maybe because of security issues, which is still something that everybody is trying to work out. But PPP is a Standard now, a new protocol compared to ones that were in development eons ago like Telnet and FTP. The HTTP protocol is crawling up. It became a proposed standard a couple of months ago (?). OTI: HTTP is still only a proposed standard? Reynolds: There is a little history behind that. HTTP was developed on the outside by the World Wide Web consortium at CERN. They went commercial with World Wide Web in I think October '92. Obviously, HTTP has taken off like crazy. So a suggestion was made by one IESG member to see if we couldn't get them to come play in our sandbox, because this would be really important for the IETF to know about on the one hand, but also we'd like to welcome them into IETF. So HTTP is developing commercially a few versions ahead of the one that is on the standards track in the IETF. The point is, they were invited and they chose to participate. There's an HTTP working group that was formed and Tim Berneys Lee (?) is one of the chairs. So HTTP is a sample of one of those outside developments being invited in. OTI: The Web went commercial in October '92? Reynolds: The World Wide Web consortium. I was at a conference in Warsaw Poland put on by EARN the European Academic Research Network. EARN's not in existence anymore (?). Robert Cailloux (?) had just come from an EC (European Communities) meeting in Brussels and he stood up in front of the whole plenary and announced the commercialization of World Wide Web was starting. In EARN in October '92. OTI: This is a good example of an outside development of a protocol which was then invited in. Who did the inviting, how did that happen? Reynolds: Eric Hauser(?) of the Netherlands was instrumental in talking to Tim (?) to get the idea of standardizing HTTP within the IETF community. Eric ran with it because he decided that was his task to go in, with the approval of the rest of the IESG. Also because Eric was in the Netherlands and at the time Tim (?) was still in Switzerland. Tim's at MIT now. OTI: There are two interesting issues here. There is the concept of developing a new protocol independently of the IETF. After all, the Internet is a world-wide structure. It's interesting to see how there could continue to be a single organization to administer it. And then I want to talk about the recent commercial pressure to develop internet technology, exemplified by Netscape, and how the IETF is coping with that. Reynolds: It's very much still growing pains with the IETF. When first started it was a bunch research guys, mostly non-profit university research based, sitting around years ago, eight or twelve of them. A year ago, December, at San Jose we had 1148 attendees and ninety working groups. When people come in who have never been to an IETF, they get overwhelmed. I had one lady from the UK who was at her first IETF. She came up to me afterwards and, very British-like, looked at me and she said, 'You people at the IETF, you all run around, you're all bumping into each other, you look like you don't know what you're doing, but you get things done'. To me, the trademark of the IETF is that we get things done. Yeah we are a bunch of researchers and developers, academic and commercial -- with an important common goal for the good of the internet. And yes we do run around and bump into each other and look like we're not organized and we don't know what the hell we're doing. A lot of the work gets done in conversations in hallways. There are other standards organizations out there that are still thinking real hard about standardizing things from twenty years ago. They go and they think real hard. We get things done. There are some doomsdayers on the other side, which I'm not a part of, who just say pretty soon the IETF just won't be there, it will self-destruct because each commercial entity will go off and do what it wants to do. OTI: Is that possible? Reynolds: I don't know yet. But instead of thinking that things are hopeless, I prefer to be hopeful. That's how I feel about the IETF and the internet community. OTI: People have been saying that the bandwidth problem was going to be a limit for along time now. Reynolds: A few years back everybody said, 'Oh my God we're running out of class A numbers. Oh my God we're running out of network numbers.' And the date that they had all scientifically worked out and predicted and ran through computers came and went. There are people out there, no matter what hats they're wearing, that sincerely want to make sure this does not happen. OTI: What is the legal status of the IETF? Reynolds: The Internet Society (ISOC) was started as a non-profit organization to be the umbrella over IETF, the IESG and the IAB. The ISOC also wanted to provide insurance, to protect members of the IESG from being sued. The ISOC provides legal help if we get in trouble. As far as liaisons with other organizations, there are documents that have been passed between the entity called the IETF and, say, the ISO. I guess they legally bind things, saying the ISO will liaise with the IETF and vice versa. OTI: Since you brought it up, can you comment on what happened between the ISO's (International Standards Organization) 7-layer network protocol model vs. TCP/IP? Reynolds: There was a religious war going on for quite a long time between the X-25 guys and 'those Americans' with the TCP/IP. 'Those Americans', that's all I ever got. Why is it my fault? I go to Europe for these conferences and it's 'You Americans'. 'You Americans'. I'm not an ugly American but when somebody starts talking at me like that, I'm going to get ugly real quick. I just came over here to give a little talk about User Services in the internet. What happened was a religious war going on until 1990, when Vint Cerf, who helped found ISOC, was the one who said it's a multi-protocol environment. Vint Cerf was giving many talks. One of his slides stated 'The Internet knows no boundaries or political boundaries. The packets cross freely we should be able to too'. OTI: The Internet knows no boundaries, the packets cross freely. This was Vint Cerf speaking for a lot of people? Reynolds: For a lot of people. This was not an Americanization of the internet because one of the four or five charter members of the Internet Society was RARE, which is the big entity in Europe for getting EC money. They run groups and task forces and work in groups very similar to the IETF. Very solidly European specific and very powerful politically. They were one of the buy-in charter members of the Internet Society, along with the CNRI, Educom, and others. OTI: So the Internet Society truly is an international organization. Reynolds: It's an international organization. The big buy-in to me was RARE: the X-500 and X-25 folks. Vint was promoting the multi- protocol environment but he could not get agreement until the Internet Society started in January 1990. That's when the religious war calmed down. OTI: Going back to the legal issues, there are some recent high visibility incidents having to do with censorship and lawsuits about domain names. Do you follow that kind of news? Reynolds: I have to. OTI: Let's consider a recent case about thegap.com, a small internet provider in Ireland. So, the GAP company, some kind of large corporation, is going after them via legal vehicles in an attempt to expropriate their domain name. What do you think of that? Reynolds: Jon Postel is really the one who has to interact with these types of disputes because they do go up to the IANA level. What we've always said is that if you've applied for it first and it was available, it is yours . If there is a conflict, the IANA is neutral. We're a central repository to administer this stuff but the two warring parties have to go and talk it out. OTI: So if the GAP wins their lawsuit in the United States, why should that be a reason for anyone in the domain name administration authority to change anything? Reynolds: That's where it's getting really touchy. We have templates and certain guidelines but it never figured into this type of stuff. We had this situation with the U.S. domain Mountain View. Where the chamber of commerce or city council of Mountain View, California wanted to have mtview.ca.us and that particular string was already taken years ago and registered and actively used for years by Marshal Rose who lives in Mountain View California. For years he has been using mtview.ca.us. Well, the city of Mountain View wanted to use that and they asked us to take it away from Marshal. It got elevated up to Jon Postel in phone calls. Jon said it has already been taken. It's been used for quite a number of years, by a particular person. It's not like that person grabbed it and is not going to use it. You go talk to Marshal Rose and convince him to let you have it. That's your problem. The chamber of commerce and the city of Mountain View kept saying, no, you go tell him. We want it because we are Mountain View blah,blah,blah. We don't and we can't ask people to just yank what they're doing. They have to go talk to the person who originally got it. Whether it's on the U.S. domain level, com level, or whatever, and then get back to the IANA or U.S. domain administrator. Jon offered other alternatives like mtvw or abbreviating it or spelling out MountainView.ca.us, Mountain hyphen. No, no, no. No room for negotiation with the city of Mountain View. OTI: Getting to the censorship issue, the Telecommunications Act of 1996 passed recently and many Web sites were blackened to protest that. Some of the provisions amount to a form of censorship. What do you think of that? Reynolds: Specifically what came up in the IETF was about child pornography and we had a BOF last July. A BOF is a Birds Of a Feather. It's not a real working group but you have all these people who are interested in venting at an IETF. You can go off in a room, call it a BOF, get a little agenda going and just have at it. We had a huge, huge turnout, but we really didn't come away with anything specific. OTI: What's huge for a BOF? Reynolds: We had about 250 people. OTI: When did this occur? Reynolds: In the Stockholm IETF in July '95. It was multicast, because it was considered very important by the IETF community. There were two issues: Did the IETF want to get in the middle of this in some form and if so, how. The general consensus was, yes the IETF is interested but we didn't know at the time how we could contribute. What happened was that the World Wide Web consortium took some of the ideas, which weren't new, coming out of that BOF. They took it, they ran with it, they're developing a protocol, let's just wait and see what they're doing. In my area, we provide documentation to teachers and trainers. It was decided that until the protocol work is finished or developed, there's nothing really my group can write about. OTI: What is your area? Reynolds: I am the area director of the User Services area. It's an appointed position within the IESG. I've got various working groups, about ten of them, that are working on anything from NOC (Network Operation Center) tools to site security handbooks to the internet user glossary. OTI: You were talking about protocols which would be similar to the V-chip idea for televisions which lets the end users decide what they want to receive. Reynolds: It's a type of filtering and it's per each person's choice. OTI: Some people would say that these protocols would enable government censorship. Reynolds: We needed an upgrade because the last telecommunications act was 1932. We needed a telecommunications act but people were jumping up and down and trying to put all sorts of odds and ends in there. We'll see what Congress does with the rest, but I have such a hard time thinking US-centric these days. Everything I think of is very global. Packets pass freely. My brain passes freely. It's not a nationalistic type thing. OTI: Some think the Internet is bigger than any one country and is a structure which can't be relegated to the confines a any single corporate structure or country. That's a new thing, because it is so accessible to so many people. Reynolds: Yeah, 'why do they get on the internet when they have human rights problems?'. Okay, here's another thing -- can you keep a country off of the internet because of their human rights problems? OTI: China has access to it. And even for countries which ostensibly have a better human rights record than others, it may be that with the internet people find out that is not always true. Reynolds: Right. Another little catch-22. OTI: What are you most proud of in your work on the net? Reynolds: I had this idea for what are called FYI (For Your Information) RFC's (ftp://venera.isi.edu/in-notes/fyi/fyi-index.html). We had at the time probably 1500 RFC's. New people coming on to the internet, whether they're Ph.D.'s in computer science, or whether they're a library science person, or an educator, or just a kid, would take a look at the index to 1500 RFC's and they'd walk away. FYI RFCs provide information such as introductions, user glossaries, handbooks, and guidelines. We've got 28 FYI RFC's since March of 1990, when we started. The FYI RFC's have been translated into different languages. When the Czech Republic became the Czech Republic, university head honchos that were putting the wires together with NSF came to me saying we are translating your FYI series into Czech. I had a gentleman from Japan who came up to me saying he was a grad student tasked with translating each FYI into Japanese from American colloquial English. OTI: Your example of how people spring up to help in the translation process of the RFCs is a good illustration to close on, as it illustrates how the internet involves spontaneous world-wide cooperation of volunteers working towards a common goal. Thank you. Reynolds. Thanks.