RE: xml:* attributes

Uche,

OK, well, I'll hold fire until I see what AF-lite might actually look like, but if it needs links (what doesn't), those at least IMHO, should be in a reference vocabulary that's reserved, or root.

And also IMHO, that vocabulary should reflect the architectural style of the web, as we've come to know it, which means (something like) what we're working on in the XML Hypermedia Community Group.

I spent a good part of my recent holiday reading Erik Wilde's book on xlink circa 2002.  It was abundantly clear in reading that text, that the general understanding of the architectural style of the web has evolved substantially since xlink was invented.  Any 2012 trip down the overly-complex and not-related-to-the-web-we-actually-work-with xlink path is effort ill spent.

It would be good to inherit (some) semantics from XML, not re-invent them, which seems to be the trend line, and over which other have also expressed concern, if I'm not misinterpreting them, which is always possible.

Anyway, count me really interested in how simple an AF mechanism can be.

Peter



________________________________
From: Uche Ogbuji [mailto:uche@ogbuji.net]
Sent: August 21, 2012 10:18
To: public-microxml@w3.org
Subject: Re: xml:* attributes

On Tue, Aug 21, 2012 at 5:55 AM, Rushforth, Peter <Peter.Rushforth@nrcan-rncan.gc.ca<mailto:Peter.Rushforth@nrcan-rncan.gc.ca>> wrote:
Sorry, pulled the trigger by accident!


Any spec that requires a transformation to be useful is too complex.  That said, I have only recently started to read up on AF.

But if you have to even so much as squint your eyes to think about what a piece of markup might mean, it is getting too complex.  Now @xml:* might appear complexer than just @*, but not by too much.  And there's no declaration required, so cut and paste and probably lots of other things "just work".  I know I personally am hoping for hypermedia someday, but whatever.  Even Atom uses xml:base after all.  It can't be that bad.

I don't think anyone is saying xml:base is bad.  James Clark gave a long list that I think covers the reasoning for omitting xml:* attributes, and none of the reasoning is so crude. Anyway even if you allow xml:base attributes you'd still need something like AF for hypermedia, since xlink:* would be prohibited under the options that have gained any serious traction.

As for AF itself, yes the way it was expressed and modeled in the SGML days was complex, but the underlying concept is simple, and that's all we want for MicroAF. You're simply transforming the content model actually used in a vocabulary to a reference one, a form, which, for example expresses link, id, base, etc.  This doesn't have to be an explicit transform. It's barely more than normal schema documentation. After all, before you use even a <span> tag in HTML, you have to learn what that tag means.  MicroAF would really just be a thin layer on top of regular schema documentation, and it doesn't even have to be an explicit layer.  View source, follow-your-nose, etc. would all still work fine.


--
Uche Ogbuji                       http://uche.ogbuji.net
Founding Partner, Zepheira        http://zepheira.com
http://wearekin.org
http://www.thenervousbreakdown.com/author/uogbuji/
http://copia.ogbuji.net
http://www.linkedin.com/in/ucheogbuji
http://twitter.com/uogbuji