--On Saturday, December 25, 2021 17:26 -0500 Phillip Hallam-Baker <phill@xxxxxxxxxxxxxxx> wrote: > On Sat, Dec 25, 2021 at 4:45 PM John C Klensin > <john-ietf@xxxxxxx> wrote: >... >> Maybe worth remembering that, if one goes back several >> centuries and in relatively small communities, the reason for >> inventing what because surnames, often starting from >> patronymics, matronymics, village or country names, >> occupations, or other attributes, all appear to have >> originated from a need to solve the same problem at a much >> less challenging scale. >> >> And, since this discussion started with an example starting >> with "@Alice", I note that a conversation that includes both >> John Levine and myself requires extra qualification and >> energy (in spite of my having a family name that is probably >> less common than Randy's and the protection of a tradition >> that forbids having two living people in the same family >> branch with the same first name). > I don't have all the answers, but I think I can provide an > improvement on the status quo. It seem to me that "better than the status quo" is rather easy, but whether it is workable and would get traction is another question. As you know, the status quo, however lousy, has a good deal of associated inertia. > One side effect of the use of patronyms was that they > reinforce patriarchy in ways that has significant impact even > today. Of course, and that is why I say "and matronymics". I was not suggesting that system (or any "traditional" one) was desirable, only that these issues are not new with the Internet. Nor was I suggesting that patronyms, if used at all, be used as names as you seem to have inferred. As attributes for disambiguation, they may still be useful because it is rare for either your birth mother's or your birth father's name to change during your lifetime (although there are many situations that makes their use impractical or impossible) In particular... > I have little difficulty making contact with my male > professional and social contacts from 30 years ago but I have > lost contact with many of my female contacts because they have > changed their names after getting married. Sometimes more than > once. But that, while obviously true, has little to do with the problem except in societies where women have no choice other than changing their names on marriage -- a custom that, as you point out, carries its own problems. I have encountered many women who picked a stable name, usually a pre-marriage one, and retained it at least for professional purposes because searching for, e.g., articles by that person is problematic if the name changes. I assume you have had similar experiences and connections. What they do, and the names they go by, in their personal lives might be different, but I don't think that is any different from the examples you have given about why one might need more than one callsign. > If people had a name that they could use for life, that would > answer that problem (or remove that advantage). Again, except for societies that require women to get married and to change their names when they do so, that is relatively easy today. The uniqueness problem is a separate issue: easier for me than for Randy or John Levine, easier for Randy than for many others, very hard for the proverbial (and actual examples of) "Joe Smith". But in none of those cases is the problem inability to maintain a name for life; it is ownership and/or disambiguation vis-a-vis others with reasonable claims to the same name. > If we are using human readable names, there is going to be > some idiot who really wants to be @thanos and a Disney lawyer > telling them they can't. Indeed. But, as Randy indirectly pointed out, if the names are not human readable (and probably human memorable) I think the system will inevitably need a dictionary or directory that maps their usual names to the unique machine readable ones, even if it uses some serial refinement model for disambiguation. Our history with implementing and deploying such directories at Internet scale has been fairly poor. > But even if that has to happen, the old binding of @thanos > will still the visible for those who want to contact the > original holder. True unless said Disney lawyer can convince you or appropriate authorities that the initial use of the name was invalid and violated their rights to a sufficient degree that the name and/or binding should be removed from the database... and can figure out a way to enforce that conclusion. > So the considerations are not the same as for > DNS names. There can be multiple bindings to the same name but > only one can be the active binding at a given time. Understood. That is an important consideration but I don't think it is the one John L., Randy, Christian, and I are poking at from somewhat different directions. Put differently, if I correctly understand what you are trying to do (and I may still not), much of what we have been talking about may be external to the problem you are trying to solve and your proposals for solving it. That doesn't make the concerns irrelevant because they may be key to whether you can get practical acceptance and uptake outside a very narrow community, but that, again, does not mean your ideas themselves are unworkable. > There are no perfect solutions but there can be better than > what we have today which is Alice has no power, only the > domain name owner has power and domain names are very > expensive to buy and to use. They are not designed for > identifying people, they are designed for identifying > organizations who own and operate computers. Here I think you are confusing several different issues. Looked at from my perspective in the 1980s, you are not talking about intrinsic properties of the DNS. That was a long time ago but the assumptions then included both the idea that charges for names higher than cost recovery were unnecessary and probably undesirable and that the DNS would be used with deep hierarchy with zones near the root operated in the public interest. The question of what the tipping point was that made were we are today inevitable is interesting but probably not useful. But neither the power relationships nor the costs are necessary properties of the DNS. In that context, the problem with identifying people is a design issue only because one would need to use sufficiently deep hierarchy --think, or example, about birth country, subdivisions as appropriate (perhaps down to geolocation), family name at birth, and then personal name at birth and a birth datetime somewhere -- to require that directory function. But is isn't a technical or, in principle, administrative problem with the DNS. Any country that felt like setting up such a system for its residents in their ccTLD and that did not consider the first priority of the DNS to be a revenue source could deploy such a system this month. Now, personally, I don't like strict hierarchies for systems like that and think it would be easy to do better. I feel even more strongly about plans that use the DNS because it is handy even though it is not a particularly good match for the problem being addressed. But, again, there is no inherent problem with the DNS that justifies what you are suggesting even though there may be multiple practical problems in the current DNS operational and management environment. best, john
