On 25 Jan 2011, at 19:44, Sean Turner wrote:
Yeah ;) Seriously, some people say that the warning in RFC 2026 about the drafts is enough (repeated below for convenience):
The problem is that as soon as code has been shared, people start to deploy it. In the open source sector, this is a real issue if the code is shadowing the draft. I hit exactly this problem with my implementation of RFC4462 for OpenSSH. There were a issues with earlier versions of that draft and with my implementation of them. A number of vendors deployed these earlier, buggy, releases and we were left maintaining backwards compatibility hacks for many years.
I don't think that there is any easy solution to this - release early, release often is always going to create these kind of conflicts, and there are problems that you only discover when implementations are produced. If those implementations are being produced as open source software, then people will deploy your development code, and expect it to not be arbitrarily broken by upgrades. So, I do feel that you are going to have to change OIDs if the meaning of those identifiers change.
Cheers, Simon. _______________________________________________ abfab mailing list [email protected] https://www.ietf.org/mailman/listinfo/abfab
