- From: Len Bullard <cbullard@HiWAAY.net>
- Date: Thu, 12 Sep 1996 08:36:24 -0500
- To: Gavin Nicol <gtn@ebt.com>
- CC: w3c-sgml-wg@w3.org
Gavin Nicol wrote: > > > - attribute values without quotes around them > > God forbid. I've seen far too many HTML pages with things like > > <A HREF=http://foo.bar> > > to be comfortable with this. Things like color specifications, > style information etc. etc. will only make errors more common. > > I would strongly recommned *requiring* that attribute values be > delimted by LIT or LITA. I second that. When doing conversions, we end up cleaning up a substantial mess because of minimizations. I assume that one application of XML will be to have a very consistent target set for translating SGML sources into. To do this, we often build ad hoc tools and these tools need simpler models. So, I prefer not to minimize and to disallow mixed content models. Our problems start where we have to allow for loosely formatted documents that a customer insists look on the screen as they did on the page. This is where we usually have to allow inclusions for font changes, pushing text around, etc. The other consideration is cost of teaching employees how to convert. The simpler the target, the better. len bullard lockheed martin
Received on Thursday, 12 September 1996 09:36:33 UTC