- From: Andrew Clover <and-py@doxdesk.com>
- Date: Sun, 11 Jul 2004 07:37:35 +0900
Ben Meadowcroft <ben at benmeadowcroft.com> wrote: > why not use the script macro syntax &{identifier} rather > than the square brackets []? 1. Because it's malformed XHTML. XML 1.0 states the only use of & in an attribute value is for a (numeric) character reference or a (named) entity reference, neither of which can accomodate this. I would be surprised if the W3C's validator didn't complain about this; the Firefox Web Developer Toolbar's one does. 2. Because it has another, incompatible meaning - Netscape 4 actually implemented the ghastly hack described in Appendix B, which is why it was mentioned there. 3. Because with the error fixup procedures of IE and other browsers, it's indistinguishable from '&{identifier};' at a DOM level. > Any comments? 4. Square brackets are still pretty bad. You can't have them in an attribute of type ID, so 'id' attributes in a WF doctype would have to be of another non-ID type, with browser hacks to make them actually IDs after all, or something, ugh. Also they cause pain for PHP users and others who like to put literal square brackets in their names; the nbzwsp-hack is in insufficient workaround as it doesn't, AFAICS, allow PHP-[ and WF-[ to be mixed. 5. Deep breath: I believe the entire idea of repeat-sections to be an poor solution to a question that can already be better answered by plain old JavaScript. It *easy* to provide a library that uses the DOM method Node.cloneNode followed by some manual or automated fixup to attribute values to produce repeat sections. Requiring this feature at browser level, with the added complexity of the clumsy square-bracket-hack and the extra-nasty nbzwsp-hack, is a Really Bad Idea in my personal opinion. -- Andrew Clover mailto:and at doxdesk.com http://www.doxdesk.com/
Received on Saturday, 10 July 2004 15:37:35 UTC