💾 Archived View for rawtext.club › ~sloum › geminilist › 001271.gmi captured on 2020-09-24 at 01:59:42. Gemini links have been rewritten to link to archived content
-=-=-=-=-=-=-
Martin Keegan martin at no.ucant.org
Tue Jun 2 22:34:08 BST 2020
- - - - - - - - - - - - - - - - - - -
On Tue, 2 Jun 2020, colecmac at protonmail.com wrote:
At this point I am honestly considering speccing that =
lines may
*only* use URLs whose scheme corresponds to an application protocol.
Permitting *any* RFC-compliant URL is just way, way too open-ended and
defeats the point of so much careful efforts elsewhere in the protocol.
I would be in favour of this, I don't want to start seeing pages that
have tags on them.
I believe I earlier proposed something along the lines of whitelisting or blacklisting of URL schemes, but I don't really have a coherent proposal.
Ultimately there is a tradeoff which must be made between the benefits of the IETF-derived approach (existing, documented standards; existing, debugged implementations; conceptual familiarity) and the drawbacks (the assumptions with which these come freighted; the "robustness principle" rhetoric; the capture of IETF standards-making by particular interests and philosophies).
I don't think the existing standards-making systems can be relied upon to produce something clean and module like a resource locator that couldn't be extended into a Turing-complete language. Therefore, if we want to have a minimalist alternative, we may need to depart somewhat from re-using IETF / W3C standards, which I would support. The WWW already exists for people who want the more complicated/ornate systems.
Mk