- From: Steinar Kjærnsrød <steinar@infostream.no>
- Date: Tue, 8 Jun 1999 13:51:28 +0200 (MET DST)
- To: Rick Jelliffe <ricko@gate.sinica.edu.tw>
- cc: html-tidy@w3.org
On Tue, 8 Jun 1999, Rick Jelliffe wrote: > From: Steinar Kjærnsrød <steinar@infostream.no> > > > But the Right Thing (TM) for the > > future would probably be for server implementors to stuff their > > processing instructions inside XML CDATA. This eliminates the need for > > parsers like tidy and others to know the gritty details about all kind of > > processing pragmas. > > I don't see why that is the right thing at all. A CDATA section just alters > the delimiter recognition rules. With server-side includes and ASP and JSP > we have a kind of tag introduced which looks, smells and tastes like a PI. But they are not valid XML PIs, which cause otherwise well formed XML/XHTML pages to be rejected by XML parsers/validators. > Why have two parsers (one for HTML/XML markup and one for <% %> markup)? We already have different types of parsers in different flavors. Stuffing proprietary PIs inside legal XML constructs would impose only small modifications and minimal overhead on parsers for those proprietary PIs. The nice thing however, is that these XML/HTML files may be procesessed by XML parsers as well. I see that as a great benefit, others may not. ## Please direct further replies to me directly, as this thread is off tpoic for the tidy list! > > Rick Jelliffe > > -- Steinar Kjærnsrød <steinar@infostream.no> InfoStream AS http://priv.infostream.no/~steinar/
Received on Tuesday, 8 June 1999 07:51:37 UTC