Re: [xml-dev] A question of necessity

[Steve Newcomb <srn@coolheads.com>]

On 04/23/2016 01:40 AM, Arjun Ray wrote:

But does it have to be HyTime?  Not as is, I don't think.  We can get
a lot of mileage out of one crucial idea from  HyTime: attribute-based
processing.  That might be enough.  Start simple and keep it simple! 

Fair enough.

The question of how simple it should be depends what problem you expected the XML Namespaces thing to solve.  The claims that were made, and that are still being at least implicitly made, are that (somehow) XML leverages the namespaces to (somehow) ensure more reliable interchange of data, where such interchange (somehow) involves multiple independently-maintained vocabularies.

The HyTime Architectural Forms (AFs) approach actually delivers on that claim, and so it's not exactly simple, even though I would argue that it's really not all that complex, either, and no more complex than it needs to be in order to meet the stated requirement. 

Aside: But my perspective may be different from the ascendant one.  Dinosaur that I am, I've never really adjusted to the idea that the cause of global information interchange is aided in any serious way by documents that don't disclose the class of document structures to which they presumably conform.  If XML isn't about global information interchange, what is it about?  That is what SGML (ISO/IEC 8879:1986 et seq.) is about, no more and no less. 

BTW, OASIS, the operator of this list, used to be called "SGML Open", with the goal of allowing industrial practices to remain consistent with that same overarching goal.  AFAIK, OASIS still actually supports that same goal, and I say "Brava!" and "Bravo!" to all its members.  Your path is not easy, but it's very worthy.

HOWEVER: If all that we're trying to accomplish is to add some sort of politically- or economically-hierarchical annotative capacity, as, indeed, XML Namespaces does do, then that's "no hill for a mountain climber".  If that's all we need or want, the AFs approach is clearly too heavy.  Nevertheless, as perhaps you intended to imply, Arjun, the syntactic tricks used in AFs for name-disambiguation would still be applicable, and if we used them, we could go back to believing that generic identifiers are opaque strings having no XML-defined internal structure.  FWIW, I think such an approach makes sense.  Or maybe you'd prefer a different way of disambiguating the political/economic context(s) in which names should be read? 

As far as I know, nothing in the existing XML standards or recs requires you to put a colon anywhere in any XML name, except perhaps for a brief gesture of obeisance at the top of the document.  If you don't believe in colonized names, don't use them.  I don't believe in them, and so I use them to the least extent possible in my work.  Your mileage may vary.

Steve Newcomb