Re: [apps-discuss] draft-ietf-iri-comparison

["Martin J. Dürst" <[e-mail]>]

Hello Tom,

On 2015/01/22 19:19, t.petch wrote:
> Larry
>
> I notice that this (expired) I-D updates RFC3986 which seems germane to
> recent discussions on this list.  Reading it, I can see why it might
> update RFC3987 or 3987bis but cannot see where it updates RFC3986.  What
> am I missing?

First, it says "3986 (if approved)", so this was conditional (of course).

Second, there's the following sentence in the introduction, which may be 
enough of an explanation:

    As every URI is also an IRI, the comparison and canonicalization
    methods also apply to URIs.

Third, I think that for this topic, 10 different people would write 10 
different stories, but the conclusion would be the same: *it depends*. 
Larry hoped that we could write a better story than RFC 3986 (and RFC 
3987, which is mostly the same story, just with some twists added). 
Also, on occasion, we had the dream to get rid of the URI vs. IRI 
distinction.

Based on the experience with that document, I'd recommend to mostly 
leave the equivalence/canonicalization/comparison story alone to save 
cycles for more immediately relevant work. Of course that doesn't 
exclude the possibility that somebody write a better story, but that can 
be done as a paper or a blog entry or an informal RFC independent of 
standards work.

Just my 2¢.

Regards,   Martin.