Repository navigation
RFC URL note has lost its context #1726
Description
As noted in #1725, this note has lost its meaning with recent changes to the URL standard:
Some language specifications reference Requests For Comments that preceded [URL], in which case the earlier RFC applies for content in that language.
This used to be a note about RFC3987, in which case it was a caution that you can't use internationalized IRIs if only RFC3986 or earlier are allowed by a format.
I suppose this is still true -- just because we talk using URL standard language doesn't make the URL standard applicable to other formats that don't reference it. But perhaps we can make this clearer and also, as @iherman noted, find a better position for it so we can make it clear it applies to the whole section.
Activity
iherman commentedon Jun 30, 2021
The only format where this may apply today is SMIL. All the others, I guess, rely on the URL spec...
If that is indeed the case, we could move this note to §1.3.4, and that is it...
mattgarrish commentedon Jun 30, 2021
I believe the only format that references the URL standard is HTML. SVG2 references RFC3987. MathML 3 mentions URIs but doesn't reference any standard and the 4 draft references RFC 3987.
But, in any case, we have to consider the extensibility of EPUB, too, even if it's an un(der)-used feature.
iherman commentedon Jun 30, 2021
Well, SVG2 references both :-)
For all new specs there is an informal requirement, afaik, to refer to the URL spec, unless there is a good reason not to. Ie, I would expect this issue to go away...
We could move the note to §1.3 as a general note.
mattgarrish commentedon Jun 30, 2021
I'm not sure about a note, but I suppose we could try to shoehorn it into a statement about the relationship to the URL standard. So maybe a new section like:
But it still feels like we're weaseling our way around a normative requirement with the last sentence, which is the crux of the original note, too.
iherman commentedon Jul 1, 2021
I do like the idea of the separate section.
As for "weaseling": the problem is that we may be facing 'living standards', which means that formats, as implemented and deployed, may be ahead of the normative equivalents. (SVG2 is probably in such a state.) If we put a MUST in the last sentence, we may artificially restrain authors. I would propose to leave it as in your proposed text.