- From: Peter Howkins <peter.howkins@oregan.net>
- Date: Mon, 30 Apr 2007 17:58:32 +0100
- To: public-html@w3.org
As a person that is likely to have to impliment HTML5 in code at some point in the future I thought I'd express a few thoughts on things that might make my life a little easier. 1) Tests Lots of them please, covering expected behaviour and edge cases. For tests of new features that have an impact on layout etc, but where no implimentation yet exists consider having a graphic of how you imagine the feature will look, so we have a reference to work against. 2) Specify failure behaviour The spec is likely to be full of wording such as "The element will contain XYZ", I'd like the spec to suggest what we should do when it doesn't. By specifying the failure behaviour as well as the correct behaviour there will be less chances for browsers to operate differently by handling failure in varying ways. 3) Version number I'm not a supporter of the idea in principle but practicality must take precedence. If the behaviour of the spec differs from previous versions of HTML then I need a way to know which method to use, a version number is as good as any other method. If failure behaviour is specified (as I suggest in point 2) and that differs from older versions, a version number or similar is also needed. 4) Tables for layout People use tables for layout, laying things out on a grid is easy. Don't feel too bad that people don't want to use css for layout, consider it a user (content designer) led trend. 5) Not too many new semantic tags If new semantic tags are being proposed that are little more than formating options consider not implimenting it, eg <foobarxyz> with a bit of style is not too far from <div style="foobarxyz"> in which the user gets to specify the style. Peter -- Peter Howkins - Senior Software Engineer - Core Team Oregan Networks UK Ltd Tel: +44 (0) 20 8846 0990 The White Building, 52-54 Glentham Road Fax: +44 (0) 20 8646 0999 Barnes, London. SW13 9JJ, United Kingdom WWW: http://www.oregan.net/
Received on Tuesday, 1 May 2007 07:58:26 UTC