Netizens-Digest Thursday, April 5 2001 Volume 01 : Number 377 Netizens Association Discussion List Digest In this issue: Re: [netz] There is a need for online discussion of new DNS NASCommmittee [netz] Comments on the NAS DNS provisional comm. appointments due 4/5/01 Re: [netz] There is a need for online discussion of new DNS NASCommmittee ---------------------------------------------------------------------- Date: Thu, 5 Apr 2001 03:53:46 -0400 (EDT) From: ronda@panix.com Subject: Re: [netz] There is a need for online discussion of new DNS NASCommmittee "Howard C. Berkowitz" hcb@clark.net wrote: >>Interesting. Because the Internet really isn't an anarchy. >> >>It's a form of servo mechanism - a system (complex system) >>that is able to learn from past behavior toward achieving >>a goal (be it a changing goal.) >I'm reminded of a few good buttons/T-shirts: > "1,000,000 lemmings can't be wrong." > "It has been proposed that a million monkeys at typewriters would > produce the works of Shakespeare. The experiment has been tried. > It is called the Internet, and no new Shakespearean material was > detected." But what do you think Shakespeare would think of the Internet? I suggest he would marvel at it :-) I'm not sure I understand exactly what you are saying. I'm saying that it was good science and research that made the Internet possible. And that we should try to look at what has been developed and how and build on it. Are you saying it was an accident? (...) >But again we come back to the issue of their being a monolithic >definition of the Internet -- which I don't really think can exist. I'm not saying its something monolithic. But I am saying that it can be understood. And that that understanding is significantly aided by understanding the research process and science at its roots. Do we disagree? >The scientific research model absolutely, positively has its role. I >don't make the distinction you do between science and engineering in >the Internet context. Even what is academically funded "Internet >research" blurs: Internet II versus Abilene. There is research on >network-enabled applications, and there is research into networking >itself. >Scientific research communities publish their open results, but >frequently have invitational workshops and other quite closed >discussion forums. Both in computing and in medicine, where I am >most familiar with research communities, there emphatically are >non-public forums. But what I have found in my studies of the development and ending of the Information Processing Techniques Office (IPTO) at ARPA (1962-1986) is that there was a change in the autonomy the Director of the Office could exert at various times, and then the Office was ended. The office had been an important government institutional form to support some of the kind of basic research that had been helpful in making the Internet possible. So the ending of the Office was a blow to that basic research, and there is a need to learn from the lessons of IPTO and to find a way to support basic research again of a similar quality. That I agree that there are forms of research now, but I haven't seen the general nature kind of research that IPTO was able to support and inspire. And IPTO provided leadership to a research community. I realize I don't know all that is going on. I do try to pay attention to what I have access to. But I also pay attention to the research that has helped to make possible the advances we have today, and that is important too as a background in understanding the research roots of these developments. I am surprised that there seems so little effort in the computer science curriculum to familarize students with the important papers that were some of the foundation of the important networking and interactive computing achievements of our times. I see you follow medicine as well as networking based research. That is a good sign. That means you make an effort to be broad, not narrow. Licklider came out of a community that studied servo mechanisms in natural and artificial systems. And I have a sense that his contributions to the creation of the IPTO and the vision that has guided networking development in good part grew from his research instincts about the human brain, as well as from his respect admiration for the power of the computer and the need for an Intergalactic network. By the way, in an earlier post you pointed to the need for human factors research expertise. I guess Licklider helped to pioneer that field. And I agree that it seems to need to be reinvigorated today, building on Licklider's vision of the need for a symbiotic relationship between the human and the computer (and I would add network). >Yet in earlier discussions, you seemed to describe Freenets as the >ultimate expression of "the Internet." I didn't say it was an ultimate expression, but it was an early means of making access to all available that was different from a commerical model. (I realize you may have a different notion of freenet ) I was referring to the Cleveland Free-Net which began in the late 1980's. I learned about it in 1988 and it was then a cooperative means of people in Cleveland working together to make dialup access to the Internet and Usenet available at no charge to the people in Cleveland using dialup modems. I got onto Cleveland Free-Net in 1992 via an Internet connection from Michigan (using telnet). And I was able to get on Usenet from the Cleveland Free-Net and also to get access to mailing lists like com-priv . The Cleveland Free-Net model spread around the US fairly rapidly and then to Canada and to a few places in Europe. For example I think there was one in Erlangen Germany at one point. >So I'm confused. There are other aspects that don't fall into either >a scientific or a public model. In a developed country, there's >little question you will have dial tone when you pick up the >telephone. In a developed country, there are quite serious questions >of Internet reliability. >Telephone networks are tightly engineered and run by rigid standards. >They are primarily run by corporations. At the same time, there is >active research into improving telephony. >Where, in your model, is the place for highly reliable IP >connectivity? Neither scientific nor "public" organizations are >geared to run production environments. >When I need to have heart surgery, I do not go to a pure scientist >whose primary interest is genetically modulated angiogenesis. I go >to a surgeon who does several procedures per week. Science is not >the ultimate goodness. >Correct me if I misperceive, but I am tending to detect a message >that "anything is better than a corporate model". Interesting. When the ARPANET as split into an operational MILNET and a research ARPANET which would communicate via TCP/IP, that was my sense of probably one of the earliest Internet's. MILNET was operational. It wasn't commercial. It seems you are equating operational with commercial and suggesting that I am denying the need for a commercial Internet. I don't equate operational with commercial. I do agree that there is probably a place for operational networks that are part of an Internet. The Cleveland Free-Net and other Free-Nets that existed in the 1990's (and there are still some) were an example of a means access that wasn't commercial but still that provided an important form of Internet access to many people. Before the breakup of AT&T, the US had the best phone company in the world with its regulated AT&T and the Bell Labs section of AT&T. AT&T was a company, but it was under certain constraints and restraints from the US government to make it possible for there to be universal access and the research component (Bell Labs) that provided the most advanced technology to make it possible to provide that universal access to telephony for people in the US. So I am not denying that there is a reason to provide the kind of service that the pre-break-up AT&T was obligated to provide to the people of the US. I am though not agreeing that this means one has to have a commercial model to provide this service, and the research that makes it possible. In fact, there are many signs that the commercial model being used to privatize the Internet cannot provide the kind of service and research that is needed for the Internet to be able to thrive in the US. In fact I have been told by one of the people who is now in ICANN that the educational network is a hole in the commercial Internet. That seems to me to be fundamentally opposed to what I have learned about the design and nature of the Internet. That seems a rationale to try to end the Internet and replace it with the kind of one big network under one political and administrative power. >>>and I want to be a bit more awake when I try to expalin >>>what I have come to understand about Internet governance >>>from the study I have done of Internet development and mechanisms >>>of feedback and the experience I have >>>had over the years online. >> >>>Basically there are systems that are built where the feedback >>>is crucial to their development and there are systems that >>>don't use feedback. The Internet is a system that is built >>>as a system dependent on feedback. >OK. Can you accept the model that business, engineering/science, and >public interest all represent sources in control systems? If so, >what is the appropriate controlling negative feedback on each? It seems to the contrary that the business model doesn't allow for the negative feedback. Rather it is a no feedback model. That seems the problem with the conception of ICANN. The business and investment folks who got themselves in power in it have no reason to hear from anyone. And their goal isn't a social goal. It is a goal that is dictated by their own self interest. This flies in the face of the social goal that Licklider articulated in the 1960s' -- of all having access to this intellectual public utility. Licklider in an article he wrote with Robert Taylor said "For society, the impact will be good or bad depending mainly on the question: Will "to be online" be a privilege or a right?" (See the end of chapter 2 of "Netizens: On the History and Impact of Usenet and the Internet" http://www.columbia.edu/~hauben/netbook) I do see that there is a need to have an infrastructure for the Internet that provides for the negative feedback to make the system be able to scale. This means a better feedback system than currently exists for users (of all strata) to be able to participate more fully in helping to identify and then solve the ongoing problems of the Internet's scaling. I proposed such a prototype collaboration to make it possible to build such a feedback system to Ira Magaziner after talking with him about my problems with what was to become ICANN. He asked me to provide an operational form to demonstrate the import of the problems I had identified with the creation of ICANN. My proposal was up at the US Department of Commerce. It was submitted before the ICANN proposal. I proposed an international scientific collaboration to review the current means of human feedback systems and to propose a way to create a more effective feedback system. The proposal is at the US Department of Commerce web site and also at http://www.columbia.edu/~rh120/dns-proposal.txt > >I have spent a few years doing research about the early mailing >lists and newsgroups on the ARPANET and early Usenet and feel >there is another important level of negative feedback that >has functioned with regard to the development of networking >and then of the Internet. But you seem to insist that the 1970s model of Internet development is the pure one and should continue to apply. In a military context, that would mean that the people that took 2nd place in the Southeast Asia War Games should have applied all of their methodology to the Persian Gulf. Or that the sexual practices of the seventies should continue in a world containing HIV. Having lived in the seventies, I'm not happy about the latter! There's no question that the current economic environment is not the classical free market, in which building value and reputation was an economic good. But any market, even a warped one, does have feedback mechanisms. Again, I keep getting a sense of anti-corporate bias. Don't get me wrong -- I've spent far too much time in corporate environments where Dilbert is real. Free the ISO 9000! Howard ------------------------------ Date: Thu, 5 Apr 2001 09:21:48 -0400 (EDT) From: ronda@panix.com Subject: [netz] Comments on the NAS DNS provisional comm. appointments due 4/5/01 Comments on the DNS provisional member appointments are due today. I planned to try to submit something, though its hard to have any sense that such comments will be worth the effort. However, I wondered if there were any thoughts on what it made sense to submit. The whole closed nature of the process and the lack of any understanding of where most of the committee appointments came from and why those people are in any way appropriate to be on the committee is, it seems a similar situation to what has happened with the ICANN situation, only there is even less light being shone on this situation. There are IETF people and technical people on the committee and that is understandable, though why these particular people isn't always obvious or if there were others who also might have been appropriate. However, the others seem from a very narrow spectrum of people. In general they seem the people convinced that the future of the Internet is as a commercenet rather than people who have a way to consider or know of the particular and unique development of the Internet and that this unique development grew out of a vision very different from any vision of a commercenet. Ronda ------------------------------ Date: Thu, 5 Apr 2001 09:43:07 -0400 From: "Howard C. Berkowitz" Subject: Re: [netz] There is a need for online discussion of new DNS NASCommmittee >"Howard C. Berkowitz" hcb@clark.net wrote: > >>>Interesting. Because the Internet really isn't an anarchy. >>> >>>It's a form of servo mechanism - a system (complex system) >>>that is able to learn from past behavior toward achieving >>>a goal (be it a changing goal.) > > >>I'm reminded of a few good buttons/T-shirts: > >> "1,000,000 lemmings can't be wrong." > >> "It has been proposed that a million monkeys at typewriters would >> produce the works of Shakespeare. The experiment has been tried. >> It is called the Internet, and no new Shakespearean material was >> detected." > >But what do you think Shakespeare would think of the Internet? > >I suggest he would marvel at it :-) > >However, I'm not sure I understand exactly what you are saying. > >I'm saying that it was good science and research that made >the Internet possible. > >And that we should try to look at what has been developed and >how and build on it. > >Are you saying it was an accident? > >(...) > >>But again we come back to the issue of their being a monolithic >>definition of the Internet -- which I don't really think can exist. > >I'm not saying its something monolithic. > >But I am saying that it can be understood. And that that >understanding is significantly aided by understanding >the research process and science at its roots. > >Do we disagree? > > >>The scientific research model absolutely, positively has its role. I >>don't make the distinction you do between science and engineering in >>the Internet context. Even what is academically funded "Internet >>research" blurs: Internet II versus Abilene. There is research on >>network-enabled applications, and there is research into networking >>itself. > >>Scientific research communities publish their open results, but >>frequently have invitational workshops and other quite closed >>discussion forums. Both in computing and in medicine, where I am >>most familiar with research communities, there emphatically are >>non-public forums. > >But what I have found in my studies of the development and >ending of the Information Processing Techniques Office (IPTO) >at ARPA (1962-1986) is that there was a change in the autonomy >the Director of the Office could exert at various times, >and then the Office was ended. The office had been an important >government institutional form to support some of the kind of basic >research that had been helpful in making the Internet possible. > >So the ending of the Office was a blow to that basic research, >and there is a need to learn from the lessons of IPTO and to >find a way to support basic research again of a similar quality. > >That I agree that there are forms of research now, but I haven't >seen the general nature kind of research that IPTO was able to >support and inspire. And IPTO provided leadership to a research >community. > >I realize I don't know all that is going on. > >I do try to pay attention to what I have access to. > >But I also pay attention to the research that has helped to >make possible the advances we have today, and that is important >too as a background in understanding the research roots of >these developments. > >I am surprised that there seems so little effort in the computer >science curriculum to familarize students with the important >papers that were some of the foundation of the important >networking and interactive computing achievements of our times. > >I see you follow medicine as well as networking based research. >That is a good sign. > >That means you make an effort to be broad, not narrow. > >Licklider came out of a community that studied servo mechanisms >in natural and artificial systems. > >And I have a sense that his contributions to the creation of >the IPTO and the vision that has guided networking development >in good part grew from his research instincts about the human >brain, as well as from his respect admiration for the power >of the computer and the need for an Intergalactic network. > >By the way, in an earlier post you pointed to the need for >human factors research expertise. I guess Licklider helped >to pioneer that field. > >And I agree that it seems to need to be reinvigorated today, >building on Licklider's vision of the need for a symbiotic >relationship between the human and the computer (and I would >add network). > > >>Yet in earlier discussions, you seemed to describe Freenets as the >>ultimate expression of "the Internet." > >I didn't say it was an ultimate expression, but it was an early >means of making access to all available that was different >from a commerical model. > >(I realize you may have a different notion of freenet ) > >I was referring to the Cleveland Free-Net which began in the late >1980's. I learned about it in 1988 and it was then a cooperative >means of people in Cleveland working together to make dialup >access to the Internet and Usenet available at no charge to >the people in Cleveland using dialup modems. > >I got onto Cleveland Free-Net in 1992 via an Internet connection >from Michigan (using telnet). And I was able to get on Usenet >from the Cleveland Free-Net and also to get access to mailing >lists like com-priv . > >The Cleveland Free-Net model spread around the US fairly rapidly >and then to Canada and to a few places in Europe. For example >I think there was one in Erlangen Germany at one point. > > >>So I'm confused. There are other aspects that don't fall into either >>a scientific or a public model. In a developed country, there's >>little question you will have dial tone when you pick up the >>telephone. In a developed country, there are quite serious questions >>of Internet reliability. > >>Telephone networks are tightly engineered and run by rigid standards. >>They are primarily run by corporations. At the same time, there is >>active research into improving telephony. > >>Where, in your model, is the place for highly reliable IP >>connectivity? Neither scientific nor "public" organizations are >>geared to run production environments. > >>When I need to have heart surgery, I do not go to a pure scientist >>whose primary interest is genetically modulated angiogenesis. I go >>to a surgeon who does several procedures per week. Science is not >>the ultimate goodness. > >>Correct me if I misperceive, but I am tending to detect a message >>that "anything is better than a corporate model". > > >Interesting. When the ARPANET as split into an operational >MILNET and a research ARPANET which would communicate via >TCP/IP, that was my sense of probably one of the earliest >Internet's. > >MILNET was operational. It wasn't commercial. It seems you are >equating operational with commercial and suggesting that I >am denying the need for a commercial Internet. > >I don't equate operational with commercial. > >I do agree that there is probably a place for operational >networks that are part of an Internet. > >The Cleveland Free-Net and other Free-Nets that existed in the >1990's (and there are still some) were an example of a means >access that wasn't commercial but still that provided >an important form of Internet access to many people. > >Before the breakup of AT&T, the US had the best phone company in >the world with its regulated AT&T and the Bell Labs section of AT&T. > >AT&T was a company, but it was under certain constraints and >restraints from the US government to make it possible for >there to be universal access and the research component >(Bell Labs) that provided the most advanced technology to make >it possible to provide that universal access to telephony >for people in the US. > >So I am not denying that there is a reason to provide >the kind of service that the pre-break-up AT&T was >obligated to provide to the people of the US. I am >though not agreeing that this means one has to have >a commercial model to provide this service, and the >research that makes it possible. > >In fact, there are many signs that the commercial model >being used to privatize the Internet cannot provide >the kind of service and research that is needed for >the Internet to be able to thrive in the US. > >In fact I have been told by one of the people who is >now in ICANN that the educational network is a >hole in the commercial Internet. > >That seems to me to be fundamentally opposed to what >I have learned about the design and nature of the Internet. > >That seems a rationale to try to end the Internet and >replace it with the kind of one big network under >one political and administrative power. > >>>>and I want to be a bit more awake when I try to expalin >>>>what I have come to understand about Internet governance >>>>from the study I have done of Internet development and mechanisms >>>>of feedback and the experience I have >>>>had over the years online. >>> >>>>Basically there are systems that are built where the feedback >>>>is crucial to their development and there are systems that >>>>don't use feedback. The Internet is a system that is built >>>>as a system dependent on feedback. > >>OK. Can you accept the model that business, engineering/science, and >>public interest all represent sources in control systems? If so, >>what is the appropriate controlling negative feedback on each? > >This is something we should discuss further as it is an interesting >question you pose. > >However, thus far the supposed effort of the US government has >been to create what seems to be a means for business governance >and that has been a real problem. > >This has not only left out the public interest and the engineering/science >community, and all the other online communities like the educational, >government, artistic, etc. But also the business model is the model >that doesn't allow for the negative feedback. Rather it is a no >feedback model. > >This seems some of the source of the problem with the conception of ICANN. >The business and investment folks who got themselves in power >in it have no reason to hear from anyone. > >And their goal isn't a social goal. It is a goal that >is dictated by their own self interest. > >This flies in the face of the social goal that Licklider >articulated in the 1960s' -- of all having access to this >intellectual public utility. Licklider in an article he >wrote with Robert Taylor said "For society, the impact >will be good or bad depending mainly on the question: >Will "to be online" be a privilege or a right?" >(See the end of chapter 2 of "Netizens: On the History >and Impact of Usenet and the Internet" >http://www.columbia.edu/~hauben/netbook) > >Also this social goal was a goal raised by the NTIA online >conference in 1994 where many people challenged the >privatization plan of the US government. > >I do see that there is a need to have an infrastructure >for the Internet that provides for the negative feedback >to make the system able to scale. This means a better >feedback system than currently exists for users (of >all strata) to be able to participate more fully in >helping to identify and then solve the ongoing problems >of the Internet's scaling. > >I proposed such a prototype collaboration to make it possible >to build such a feedback system to Ira Magaziner after >talking with him about my problems with what was to become >ICANN. He asked me to provide an operational form >to demonstrate the import of the problems I had identified >with the creation of ICANN. > >My proposal was up at the US Department of Commerce. It >was submitted before the ICANN proposal. > >I proposed an international scientific collaboration to >review the current means of human feedback systems >and to propose a way to create a more effective feedback >system. > >The proposal is at the US Department of Commerce web site >and also at http://www.columbia.edu/~rh120/dns_proposal.txt > >You might be interested in taking a look at it and commenting >on it. >> >>I have spent a few years doing research about the early mailing >>lists and newsgroups on the ARPANET and early Usenet and feel >>there is another important level of negative feedback that >>has functioned with regard to the development of networking >>and then of the Internet. > >>But you seem to insist that the 1970s model of Internet development >>is the pure one and should continue to apply. > >You didn't say the problem with open architecture and the >creation of the tcp/ip protocol. Several problems, which can equally be considered opportunities. First, there never was a formal TCP/IP architecture, at the level of the OSI or B-ISDN architectures. There were general principles, above all, Dave Clark's "be liberal in what you accept, be conservative in what you send" (which my European friends tell me often has a completely opposite meaning in their political contexts). The IETF's guiding principle has always been (I think again from Dave Clark) "We don't believe in kings, presidents, or voting. We believe in rough consensus and running code." There were other, more technical architectural guidelines. Brian Carpenter reviewed them in RFC 1958, and I think has a more recent version. Many of these principles served us well, but are often the exception rather than the rule. They include: end-to-end significance of IP addresses primary higher-layer intelligence in end hosts local vs. remote assumption (if I'm on the same subnet, I have layer 2 connectivity, otherwise, I need to go to a router). a central interprovider core So when you say architecture, I'm confused. Most of the scientific/engineering architectures with which I'm familiar are very formal. > >Nor do you look to see what this model has created and what >can be learned from the achievement. A great deal has been created. But the participation models have changed. In the first IETF meetings, all the participants could sit in one room. At the last IETF in Minneapolis, drawing from memory, there were 2000-odd participants in 140 sessions. Presumably due to the economy, participation was down from the last couple of meetings - -- and I am not ignoring that the real work gets done on mailing lists. It may sound silly to some, but one of the values of the face-to-face meetings is associating a face and a voice with an email address. > >Instead it seems you are trying to discredit the architecture >by claiming it is from the 1970s. Again I ask, WHAT ARCHITECTURE? Architecture is not the same as the development process/model. > >Isn't this a bit like trying to discredit the US by saying that >it grew from a 1770's model (the Declaration of Independence) >or an 1780's model (the Constitution). > >If you have a critique of open architecture and the creation >of TCP/IP, you can make it. But it seems that to give the date >that an advance happened as the means to discredit it >isn't very helpful. TCP/IP did not spring forth on a particular date. It's unclear, in this context, if you are equating TCP/IP (two specific protocols) to an architecture. That equation is made informally, but it really doesn't hold. To go back to the overly specified OSI architecture (which still had value), it would be equivalent to speak of TP4/CLNP as equivalent to the OSI Reference Model--which would be incorrect. > >(...) > >>There's no question that the current economic environment is not the >>classical free market, in which building value and reputation was an >>economic good. But any market, even a warped one, does have feedback >>mechanisms. Again, I keep getting a sense of anti-corporate bias. > >The market model seems to be an equilibrium model, not a dynamic model. > >It's the feedback model, however, that allows for internal and external >changes and adjustments when needed. It isn't a feedback model >that allows for a changing goal and for feedback to understand >how to reach that goal. > >And its the feedback model that was the basis to create and >develop the Internet. And its the feedback model that is needed >to determine how to scale it. I'll need to get a better understanding of your model. Is it plausible that there are multiple coexisting models -- perhaps some type of market model that drives investment in commercial infrastructure, and feedback model for protocol development? > >That is the findings of my research thus far. > >There is a need to understand the implications of this. > >But the NAS DNS committee flies in the face of these findings. > >>Don't get me wrong -- I've spent far too much time in corporate >>environments where Dilbert is real. > >I can understand. I have had my experiences as well :( > >And I continue to have them with the research agencies that seem >determined to impose a market model on the Internet rather than >to provide the needed support for the work to understand the >model that is relevant to the development of the Internet. Ah...but which Internet? > >(I just had a bad experience with the NSF :( ) And you're an academic? My memories of college was that it was like government, but without the efficiency and responsiveness. :-) > >But the importance of the research is it helps to understand >what is needed and that will find a way to have its effect :) > Looking at it now. > > > Netizens: On the History and Impact > of Usenet and the Internet > http://www.columbia.edu/~hauben/netbook/ > in print edition ISBN 0-8186-7706-6 ------------------------------ End of Netizens-Digest V1 #377 ******************************