On Wed, Dec 15, 2010 at 11:43:31AM -0500, Worley, Dale R (Dale) wrote: > Given that, I would expect that a URL that is "demonstrably the most > stable reference available" would suggest that the contents of the > URL at any date near the publication date of the RFC "should" yield > the same contents. If an RFC references a URL that is unstable > enough for us to notice and complain about it, the URL should not > have been used in the first place. But as I tried to explain in my note, _it doesn't matter_ whether the URL is stable during the time you happen to be checking. When you make a reference to a book, magazine article, or newspaper column, you always include information like the series, edition, publication date, and so on. This isn't for decoration; it's there so that later, if someone wanted to look the thing up again, they could. Such a desire could happen years later. (Irrelevant aside: In a previous career I was regularly reading things about eighty years old, but I could still follow the citations back to the source that my author had been reading. In one case, this turned out to be really significant because my author happened to be reading a really bad translation, and a fundamental part of his confusion was easily explained once you realised that the translation he'd been working with sucked.) In the age of the web, this ability is almost certainly lost. We can, however, at least know from the citation when the information was current. That knowledge alone could turn out to be useful. Suppose, for instance, that it's necessary to refer to evidence gathered by the foobar project, and that the foobar project only publishes its stuff on its website. The URL of the website is certainly the most authoritative source, and therefore the most stable reference available, because it happens to be the only such reference. Now, if you happen to know that version 1 of the foobar project (call it wikifoobar) was maintained until 12 December 2010, when a fork happened and version 2 (call it openfoobar) competed with version 1 for being the "correct" version, then knowing that the reference was generated on 30 November 2010 will give you a pretty good clue about what version of the foobar project you need to go chase, even years later. RFCs form an archival series, and we need to use archival-series rules for references in them. A -- Andrew Sullivan ajs@xxxxxxxxxxxx Shinkuro, Inc. _______________________________________________ Ietf mailing list Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf