- From: Liam R E Quin <liam@w3.org>
- Date: Thu, 06 Sep 2012 01:47:48 -0400
- To: James Clark <jjc@jclark.com>
- Cc: public-microxml <public-microxml@w3.org>
On Thu, 2012-09-06 at 11:48 +0700, James Clark wrote: [...] > It's about as far from good and clean as anything can be. Awesome, I'll take two then please! :-) More to the point, when an elegant model of the world doesn't fit the world, you can change the model or change the world or live with an ill-fitting model. I think I prefer John's more pragmatic way to view two styles of comment here. A more radical alternative would be to throw out <!-- ...--> comments and only have <?...?> comments, but that suffers from some of the same problems as the PostScript %% rule in regards to robustness. The ESIS/MSIS or (in XML terms) XML-as-XML vs XML-as-text distinction is one of many possible distinctions. In fact, most XML editors (and some SGML editors) work on an extended abstract model, and do lose distinctions like whether attributes had single or double quotes, or whether you used SGML minimization, or spaces and indenting inside tags (sometimes outside tags too). Some preserve entity references (e.g. Serna is awesome at that) and some don't. XML applications aren't required to pass much of anything back to the application, and are free to strip PIs. But either PIs should be in the data model and can occur everywhere, or they're treated like comments and applications that want them have to find their own way to keep them. I don't think hacks like making them properties of the top-level element and moving them around are a good idea. Long reply, sorry, I should try sleeping :) Liam -- Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/ Pictures from old books: http://fromoldbooks.org/
Received on Thursday, 6 September 2012 05:48:26 UTC