Re: [apps-discuss] Retroactive application of draft-ietf-appsawg-uri-scheme-reg - comprehensive review

[Barry Leiba <[e-mail]>]

Just to one point:

> My reason is that there is software out there that is written to accept
> *anything* that is a valid URI, and legitimately expect such strings to
> confirm to RFC3986 syntax (not withstanding Sam Ruby's efforts to document
> variations between systems about what they accept as a valid URI).

Absolutely.  And the URNbis document maintains that, but keeping URNs
syntactically URIs.

> If URNs
> are to be usable as URIs in such circumstances, I think it's important that
> they follow the rules, including those concerning the role of fragment
> identifiers.

Here's where I don't agree.  If URN fragments (f-components, or
friggle-fraggles... the bit after the "#") are semantically somewhat
different to what 3986 says about fragments in locators, because
that's what semantically makes sense for URNs, why does that stop
anyone from continuing to use them in places where URIs are
syntactically required?

Barry