--On Friday, 02 November, 2007 10:14 +1300 Brian E Carpenter <[EMAIL PROTECTED]> wrote:
> One idea that was floated a couple of years ago, as part of a > one-level > standards track, was to retain the register of implementation > reports > (http://www.ietf.org/IESG/implementation.html) and mark the > entries > that have been approved by the IESG. The RFC index could then > point to > approved implementation reports, without any formal > "promotion" needed. Back in the days when words like "simple", as in "Simple foo protocol" meant what one would expect, I think this would have been a very useful alternative. Today, most of our protocols have many features, options, and characteristics. I suggest that only rarely is every one of them implemented, comprehensively and faithfully, in every implementation that might appear in an implementation report. In addition, rarely has an initial specification of a protocol turned out to be completely clear and comprehensive. Consequently, a properly-written and accurate implementation report is likely to be a mixture of what was fully implemented without problems, what was partially implemented or ignored (with explanations), and what was implemented in some particular way with other possible interpretations. The only way to take those reports and produce a clear description of what the standard should be, as reflected by the combination of the initial specification and practice experience, involves some sort of consensus document. That document could be either a revision of the original or an extended description of which parts of the original were still relevant together with corrections and clarifications as needed. A mere collection of implementation reports would be significantly helpful to a reader trying to figure out what the standard should be iff it was associated with an Applicability Statement that did the latter job. FWIW, I think our experience has been that documents suitable for publication as Draft or Full Standards are easier to write, and written more often, than Applicability Statements of that type. john _______________________________________________ Ietf mailing list Ietf@ietf.org https://www1.ietf.org/mailman/listinfo/ietf
