W3C home > Mailing lists > Public > spec-prod@w3.org > July to September 2012

RE: Is there a way in respec to hyperlink to external refs?

From: Larry Masinter <masinter@adobe.com>
Date: Sat, 1 Sep 2012 17:33:15 -0700
To: Robin Berjon <robin@berjon.com>
CC: "spec-prod@w3.org Prod" <spec-prod@w3.org>, "Joe Hildebrand (jhildebr@cisco.com)" <jhildebr@cisco.com>, "Henry S. Thompson" <ht@inf.ed.ac.uk>
Message-ID: <C68CB012D9182D408CED7B884F441D4D1E2DEF25CE@nambxv01a.corp.adobe.com>
My real point was that the "bibliographic database" you're constructing functions like a URN namespace -- there's an authority that maps from short names to longer bibliographic references.  Everyone relies on the authority to maintain the database, fix errors in it, but not change things that aren't errors. There's an implicit long-term maintenance issue which you haven't addressed.

There is a URN namespace for IETF publications specifically for the reason that it wants to reference not a particular instance of a document but rather the long-term promise by the IETF of what "RFC 2324" means, independent of location, hash, format.

-----Original Message-----
From: Robin Berjon [mailto:robin@berjon.com] 
Sent: Thursday, August 30, 2012 3:49 AM
To: Larry Masinter
Cc: spec-prod@w3.org Prod
Subject: Re: Is there a way in respec to hyperlink to external refs?

On Aug 29, 2012, at 21:41 , Larry Masinter wrote:
> Define a URN scheme for spec references. In a world where URNs work, the source format would be better than the destination.

Shane wrote:
> Your "scheme" approach is very much like a CURIE - and I like that. I bet we could even do it in a way that was CURIE conforming... declare it as a 'prefix' and it would be dereferenced in the follow-your-nose manner that you would expect from RDF
> 
> @prefix="spec: http://www.w3.org/URI_to_respec_bib/" somewhere in early in the spec, then transform the  references.

The one and only goal for ReSpec is to make the job of editing a spec as simple and straightforward as possible. This means having as little boilerplate as possible, as little syntax as possible, adherence to DRY and DWIM, and whenever possible a syntax that's as mnemonic as possible. In other words, it tries hard (and will keep trying harder) to ruthlessly be an authoring format.

Irrespective of their merits in broader, distributed, loosely coupled systems that CURIEs and URNs may have and that can make them well suited to environments in which authorial needs are balanced by other considerations, I don't think that they make the cut in the face of those design principles :)

-- 
Robin Berjon - http://berjon.com/ - @robinberjon
Received on Sunday, 2 September 2012 00:33:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 2 September 2012 00:33:57 GMT