- From: Bruce Miller <bruce.miller@nist.gov>
- Date: Mon, 06 Feb 2012 09:27:38 -0500
- To: William F Hammond <hammond@csc.albany.edu>
- Cc: Neil Soiffer <NeilS@dessci.com>, "www-math@w3.org" <www-math@w3.org>
On 02/05/2012 06:18 PM, William F Hammond wrote: > Hi Bruce, > >> Indeed MathPlayer seems to be accepting it both ways >> (maybe it always did, and my memory's going), >> as does firefox, so the prefix-less way looks like the >> best solution... > > I've always thought the verbose way (writing out the xmlns on > each<math> element) was best. > > Among the reasons: ... I think I'll be a curmudgeon when I grow up --- I seem to have an affinity for the unpopular choices, and am grumpy about them: I prefer xml over sgml and certainly html5; prefixed namespaces over non, heck namespaces at all!! But your list is pretty convincing; I'll go this route (but I'll still grumble about it). Thanks, Bill; bruce > 1. A larger class of processors can deal with it. > > 2. I want to view xml for documents (as opposed to EDI) as as a > category that is a subcategory of SGML. (And there continues to be > "political" resistance to the use of xml namespaces in author-level > xml document types for documents.) > > 3. With verbose use of xmlns and a few other conventions, it's > possible to generate xhtml+mathml document instances that require only > a couple of revisions near the top to become correct text/html > instances of html5. > > 4. I have a private local use sgml definition for a profiled subset > of html5, text/html serialization, as an sgml document type. > > BTW, Henri Sivonen's html5 online validator is found at > http://html5.validator.nu/ > > -- Bill >
Received on Monday, 6 February 2012 15:41:37 UTC