- From: Bjoern Hoehrmann <derhoermi@gmx.net>
- Date: Wed, 20 Jul 2005 01:37:22 +0200
- To: Karl Dubost <karl@w3.org>
- Cc: www-qa@w3.org
* Karl Dubost wrote: >You mentionned Elements and attributes, do you limit your thoughts to >XML languages? It's a tricky thing. I'm trying to think about >something which is a marker (not giving it the name of elements) in a >language, etc. Yes, or more precisely, I limit this to classes of markup languages. As I pointed out, there are very different kinds of markup languages. One language might make heavy use of attributes while others are virtually free of attributes, or the language might model something else, e.g., SVG defines properties for many things which you can use both as XML attributes and in style sheets, and XAML models object hierarchies and the "elements" are defined in the API specifications. There are also many different uses of the specifications, as you point out in your brief reviews, if you don't know much about a format and look straight at some element definition, you rather want to have some introductory text that explains what you are looking at than a quick overview of the features of the element an experienced user might like to have when looking for the name of a feature he can't remember. I am more looking for things many of the specifications do in similar but different way. You found for example that some specifications have schema fragments as the first thing in such a definition and others put them at the end. Or many of them use some pseudo-code instead of schema fragments for similar purposes. Yet specs like P3P, WSDL, XSLT, and Relax NG, which all do something like that, use different formats. Some specifications like HTML 4 and XSL 1.0 list attributes defined elswhere for each element, SVG 1.1 on the other hand does not even though the markup language is, from a specification point of view, not very different. So what I am looking at are questions like, is it useful to list attribute-defined-elsewhere for each element or should this be done in some index -- or not at all? There are many other differences, Atom for example uses a lot of RFC 2119 keywords to define how elements may nest, which attributes are required, etc. while DSig Core explicitly does not do that, or WSDL and WS-Chor, define the language in terms of attribute information items, element information items, etc. while others do not refer to such information items at all. There are often good reasons for such differences, yet I think there is potential and need to reduce them. -- Björn Höhrmann · mailto:bjoern@hoehrmann.de · http://bjoern.hoehrmann.de Weinh. Str. 22 · Telefon: +49(0)621/4309674 · http://www.bjoernsworld.de 68309 Mannheim · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/
Received on Tuesday, 19 July 2005 23:37:35 UTC