--On Monday, December 23, 2024 15:34 +1300 Brian E Carpenter <brian.e.carpen...@gmail.com> wrote:
> On 23-Dec-24 15:11, Salz, Rich wrote: >>> And if the email bounces, that itself if a sign. >> >> Not necessarily. Someone might have reported an errata, heard >> nothing, forgot about it, and retired or changed jobs. > > Of course. But my hypothesis is that if we deal with the "low > hanging fruit" of rather old reports that are just sitting there, > by arbitarily closing them, then any that actually matter will get > re-opened when they next bite somebody. That's exactly what happens > with git issues. As long as re-opening an issue is easy, we don't > risk much damage by closing old ones. For the majority, probably the vast majority, of cases, sure. If a problem is reported with a core document and the report discarded because no one got around to dealing with it, and that problem is in any way significant, it will be reported again (and/or the earlier report/issue reopened). For those cases, it is even reasonable to assume that the absence of additional reports could be an indication of lack of significance. The advantage of arbitrarily closing them would then be to see if others come up with new reports of the same problem: if the older reports are listed as open, someone else noticing the problem might review the existing reports, conclude that it has already been reported, and move on (something the current tooling, IMO, encourages). That is, in turn, an argument for dropping unprocessed reports far sooner than waiting for them to get "rather old". _However_, I continue to worry about those cases in which there is a subtle issue with a document whose subject matter lies outside the expertise of a large fraction of the IETF community. Suppose someone who is expert (and not a very active IETF participant for other topics) notices a problem and reports it. There is a significant risk that they will take non-response or non-action for a rather long time as an indication that the IETF does not care and, consequently, go away and deprive us of that expertise for that document and any future related ones. If we take the next step you suggest and say "nothing has been done about this, even a response, for a rather long time, so we are discarding the report", there is, IMO, an even more significant risk that they will not only walk away, but tell their colleagues who are working on that subject matter. That would likely have two effects: (i) to reduce, perhaps to zero, the odds that the problem would be reported again in the future and (ii) to damage the IETF's credibility in that area of work (for both existing and future documents). That might be an argument for being more careful about work we take on and including the question of whether the IETF will be able to maintain the document long-term, but that is a rather different subject, especially when non-IETF streams are considered. > Having spent a little time staring at errata.json, I don't think > there's even much wrong with the existing data model - the problems > are in how we use it. Define a good policy and the implementation > will follow. I probably agree, but that position, turned around slightly, is just another argument that what we have been doing (or not doing) is rather poor. Unless we have a way to change that within the framework of the existing model, it is an argument for changing the model ... if only to reinforce the idea that we really intend to do things differently. best, john -- rswg mailing list -- rswg@rfc-editor.org To unsubscribe send an email to rswg-le...@rfc-editor.org
