On 01/08/2012 18:16, Juergen Schoenwaelder wrote: > On Wed, Aug 01, 2012 at 09:42:44AM -0700, Tassos Chatzithomaoglou wrote: >> Wouldn't it be an option to have all applications & systems accept >> as input both formats, but only give as output the new one? >> i.e. browsers already rewrite URIs. > > There are other standards that currently state that % is the canonical > format for zone identifier separation. We would have to revise those > standards (and all standards using those standards) and for some of > them this is not just a simple update due to the way versioning works. > As such, a new '-' notation won't be accepted by certain interfaces > for a long time. In other words, we cause problems (perhaps for 10-20 > years) where there are currently no problems. And the question is > whether this price is justified to address the zone identifier in URI > issue. > > For me, the priority is this: > > a) Check seriously whether %en1 is really not acceptable since there > really is no ambiguity. Using this notation is what the user wants.
Of course we have done that, twice (a few years ago and this year). Dead end. > > b) If a) is indeed not possible, simply apply the URI escaping rules, > that is %25en1. This is consistent with URI escaping (which might > happen on other parts of the zone index as well). Provide advise that > URI parser implementors may accept %en1 (when it is unambiguous) and > turn it into %25en1. (Many browsers already do this kind of thing > today for other characters that need escaping.) This, to my understanding, was already rejected by 6man some months ago, with the conclusion that a new separator is needed. It's not my job to make the consensus call, however. > > I see no value in introducing a new separator. The value is providing a long-term path to cut and paste. Otherwise, I assume we would indeed choose the %25 approach. Brian > > /js > -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------