- From: James Clark <jjc@jclark.com>
- Date: Wed, 5 Sep 2012 09:00:16 +0700
- To: John Cowan <cowan@mercury.ccil.org>
- Cc: public-microxml@w3.org
Received on Wednesday, 5 September 2012 02:01:04 UTC
> > > > c) XML declaration > > I think that's clearly unnecessary, but point for point: > > 1) The only encoding allowed is UTF-8. > > 2) Versioning this kind of formats just doesn't work, as I have > discovered the hard way. > > 3) There is no notion of a non-standalone document. Agreed. My thinking was that if XML serializers output an XML declaration, then the XML they output won't be MicroXML anyway, which reduces any practical benefit we might get by allowing > in attribute values. I think my preference is now not to allow > in attribute values. After all, XML disallows < in attribute values only for symmetry with content (SGML IIRC allows < in attribute values): it doesn't make much sense to disallow < in attribute values because it's disallowed in content, but then allow > in attribute values even though it's disallowed in content. James
Received on Wednesday, 5 September 2012 02:01:04 UTC