- From: Eric J. Bowman <eric@bisonsystems.net>
- Date: Wed, 13 Jul 2011 14:39:51 -0600
- To: Larry Masinter <masinter@adobe.com>
- Cc: "karld@opera.com" <karld@opera.com>, "nrm@arcanedomain.com" <nrm@arcanedomain.com>, "www-tag@w3.org" <www-tag@w3.org>, "ndw@nwalsh.com" <ndw@nwalsh.com>
Larry Masinter wrote: > > > I suppose there's some design pattern "SIFR" which produces pages > > that cannot be expressed in polyglot? > > > > > > > *Why* can't "SIFR" be served as application/xhtml+xml? > > > > > See points 1 and 5, here: > > > http://www.w3.org/TR/xhtml1/#C_13 > > > HTML normalizes to uppercase, XML doesn't; SIFR relies on uppercase > > normalization, breaks otherwise. > > It seems like this could be fixed by XML, HTML, or CSS. If the > price of convergence of HTML and XML is making some change like this > ("xml2" = "xml except it normalizes to uppercase" ?) > I tried doing SIFR as application/xhtml+xml some years ago; had I thought the reasons it didn't work would come up, I'd have written them down. So I forget exactly where the breakage occurs, only that it isn't easily fixed. IOW, don't quote me on this, I was hoping someone else would chime in with the nitty-gritty. > > Why is normalization part of the XML definition at all? Isn't this > just an odd artifact of CSS? > XHTML requires element names to be lowercase, not XML. In retrospect, it would've made more sense for XHTML to require uppercase elements, as this is how HTML parsers are supposed to normalize (and why Firebug, etc. present the DOM using uppercase regardless of HTML/XHTML). This would have led to lots less script breakage for us application/xhtml+xml die-hards. -Eric
Received on Wednesday, 13 July 2011 20:40:28 UTC