On 02/04/2013 12:45 PM, Jouni Korhonen wrote: > >> On 02/04/2013 06:42 AM, Rémi Després wrote: >>> In this respect, changing now the IPv6 specification for hosts that >>> configure IIDs having u=1, although no serious need has been >>> identified, and this specification has ben used, would be in my >>> understanding very counterproductive: stability of already used >>> specifications is important for their success. >> >> I strongly disagree. For many reasons, including: >> >> 1) What's broken, is broken. And if it's known to be broken, it should >> be fixed. > > It is awkward but not broken imho, so in this respect I would think twice
FWIW, my comment was in response of the general statement "stability of already used specifications is important for their success". Actually, I'm not opposed to "stable specifications", but about resistance to change/fix/improve those things just because to maintain the status quo. > changing RFC4291. However, I do agree Randy about magic bits and the pain > involved. Actually, a lot IPv6 business recently has been around adding > magic to IPv6 addressing, which I kind of find uncomfortable.. Not sure what you mean. For instance, draft-ietf-6man-stable-privacy-addresses doesn't change the architecture, but simply is an alternative algorithm that fits in the existing architecture. Cheers, -- Fernando Gont SI6 Networks e-mail: [email protected] PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492 -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
