Pekka Savola <pekkas@xxxxxxxxxx> wrote: > > But as the vast evidence makes it clear, most ISPs do a very lousy job > of deploying networks *properly*, causing harm to the Internet as a > whole. Wouldn't it be somewhat in the IETF's business to try to give > advice (using BCP and Info documents) how to do it better? This started out as a response to Pekka. And a response to Pekka is still very important. But that will have to wait for another message, because there's too much background which must be covered first. >> http://www.alvestrand.no/ietf/chair/change-status.html In that document, Harald refers to: http://www.alvestrand.no/ietf/chair/ietf-mission.html which seems to state a "mission" for IETF: " The purpose of the IETF is to create high quality, relevant, and timely " standards for the Internet. This is one of those "motherhood and apple pie" statements we've been conditioned to ignore. Much though I like Harald, I don't think it's a good statement of our purpose, least of all mission. But first, let Harald explain: " Note that this clearly positions the IETF primarily as a standards " development organization. I doubt the wisdom of "positioning" the IETF "primarily" as a "standards development organization". Clearly we do develop standards; I'll even agree that there seems to be consensus to learn to do so better. But I don't believe that has ever before been our "primary" purpose. (Surely, if it were, we'd have put more effort into doing it well.) And, most of all, I don't believe it wise to try to "position" the IETF as anything other than a pretty chaotic group of people. No, strike that -- I should have said, many pretty chaotic groups of people. "Positioning" the IETF makes "herding cats" look easy! " There are other activities in the IETF; but if the IETF does not do its " core mission, all else will quickly fade. I contend that the IETF hasn't been doing a good job of "creating high quality, relevant, and timely standards", but the Internet still works. (So I must disagree with Harald here.) " This is intended to be an ordered list of characteristics. Timely " standards of low quality or that are irrelevant will not serve the " Internet's or the IETF's needs. Here we get to the crux of the matter. "Timely" standards are crucial to the operation of the Internet. Witness Microsoft. They have a wildly successful business plan based upon "timely" standards, of whatever quality and relevance. Thankfully, Microsoft has competitors. These competitors realize that "high quality" will avail them nothing if they don't have a "timely" alternative to the Microsoft standard. (This may be why so much gets implemented off of "Internet drafts".) "Relevance" is obvious only in hindsight. Even in hindsight, it's obvious that relevance -- in the eye of the beholder -- is a slippery thing. The _perception_ of relevance must be built up -- usually after a "standard" is fully described" -- in the eyes of potential users. "Relevance" simply cannot precede "timeliness". " This leaves open the very interesting and difficult questions of how to " measure quality, relevance, and timeliness. The IETF has identified " interoperability, security, and scalability as essential, but without " attaching measurements to those characteristics. "Interoperability" is a long-standing tradition. "Security" is a relatively recent buzzword -- suffering from all the problems which come with buzzwords. "Scalability" deserves a place next to "interoperability" as a long-standing tradition; but I'm fearful that -- as individuals -- we're attaching too many different definitions to it. " It is important that this is "For the Internet," and does not include " everything that happens to use IP. IP is being used in a myriad of " real-world applications, such as controlling street lights, but the " IETF does not standardize those applications. That concludes Harald's section on "The IETF Mission". Though it's pointless to try to substitute my wisdom for that of the IESG, I will suggest some food for thought: 1) The IETF has always chartered Working Groups. Doesn't that belong fairly high in our priorities? 2) There is a long tradition of waiting for grassroots interest and gathering interested persons to hammer out an approach which may reach "rough consensus and working code". Is this no longer a good approach? 3) The IETF has always had a mechanism for appropriately publishing individual viewpoints. Is this no longer important? 4) "Timeliness" is a difficult nut to crack. Wouldn't we be better off creating several new paths to improve timeliness instead of placing blind faith in an ability to "charter in" timeliness to a standards process designed around a slow consensus process? -- John Leslie <john@xxxxxxx>