Here's a small example on how to pass W3C validation while enjoying Dojo's de facto standard of mixing-in new attributes. This approach leverages XHTML1.1 Modularization http://www.w3schools.com/xHTML/xhtml_modules.asp.
W3C Validating XHTML: http://yadtd.info/dojo-release-1.1.1/dijit/demos/formw3c.html
Original HTML: http://archive.dojotoolkit.org/dojo-2008-07-01/dojotoolkit/dijit/demos/f...
Differences (HTML formatted): http://yadtd.info/dojo-release-1.1.1/dijit/demos/formw3c.diff.html
Differences (diff -u): http://yadtd.info/dojo-release-1.1.1/dijit/demos/formw3c.diff.txt
Example XHTML1.1 Module: http://yadtd.info/DTD/beta/yadtdeUIWw4.mod
Example XHTML1.1 modular DTD plus the new module (see end of file): http://yadtd.info/DTD/beta/yadtdeUIWw4.dtd
Example XHTML1.1 modular DTD plus the new module -- as a self-contained DTD (also see end of file): http://yadtd.info/DTD/beta/yadtdeUIWw4-flat.dtd
Best regards,
-Tim
ps: I'm not suggesting any need for validation -- just explaining one possible workaround in case it ever comes up.
that's really great Tim.
that's really great Tim. I'm sure if it were absolutely complete dojotoolkit.org would be willing to host the DTD def ... I jokingly suggested we should create a whole new one and name it DHTML/1.0 ;)
... let's start it at
... let's start it at version 5.0 for added benefit. ;) Any thoughts about what an absolutely complete DHTML/5.0 Dojo DTD might look like?
The most interesting discovery here seems to be that XHTML1.1 DTD Modules are fairly easy to work with, but only after the initial setup is established. Compare the .mod above to the rest of that DTD infrastructure!
Anyway, if I had to produce an absolutely complete DHTML/5.0 Dojo DTD, I'd probably just ask for an absolutely complete sample-set covering each logical markup unit. Maybe an xhtml folder, with a bunch of test_MarkupThingy_N.htmls that can be boiled down into DTD .mods. But, I digress... my goals were actually met by simply generating yet another DTD as often as needed. ;)
I'd bet the api parser could
I'd bet the api parser could pull it off ... no work involved. any element can have any attribute. it'd be a huge DTD.
Hmm, it's not too bad
Hmm, it's not too bad actually (or maybe it's worse) -- you triggered my curiosity which led to a quick reconnaissance mission:
dumpdefacto.sh: tidy --doctype strict -e -q -asxhtml $1 2>&1|fgrep proprietary|sed -e 's/^.*: //'
find . -name \*.html -exec dumpdefacto.sh {} \;|sort|uniq > dojo111defacto.txt
http://yadtd.info/dojo-release-1.1.1/dojo111defacto.txt
So it looks like across the entire sample set contained in 1.1.1, including templates, there's 466 distinct nodeName, attribute combinations represented.