- From: David Carlisle <davidc@nag.co.uk>
- Date: Tue, 14 Aug 2012 00:07:22 +0100
- To: John Cowan <cowan@mercury.ccil.org>
- CC: Dave Pawson <dave.pawson@gmail.com>, MicroXML <public-microxml@w3.org>
On 13/08/2012 22:12, John Cowan wrote: >> what would not be good at all would be if that syntax made an >> element or text or some other incompatible thing at the data model >> level. > > That would be absurd. Not at all. xquery for example uses the syntax <!-- --> for something that isn't a comment. Why should it be absurd for micro-xml to do so if the spec doesn't specify that is not the case? If micro-xml spec goes out of its way to stress that it only has syntax compatibility with xml then a micro-xml system could report <!-- --> as an element and <zzz/> as text. Do you really want that to be conforming? (I may agree with you that it is absurd, but that doesn't mean that the spec should not specify it) > >> If you push for compatibility at the level of syntax only you get >> things like html parsing of <foo/> which is accepted syntax but it >> is a start tag not an empty tag. I fail to see why that is useful >> (in general, or in html) > > Because people who know too much XML got in the habit of writing > <br/> instead of proper HTML <br> as an empty-tag, and so HTML5 > legitimized it. > Yes but html could have legitimised it by saying /> meant empty (probably with a few security/legacy restrictions such as <script>) rather than saying that the / is ignored and <sss/> parses as <sss> always. David
Received on Monday, 13 August 2012 23:07:46 UTC