[tei-council] whither HTML of Guidelines and @lang

[Sebastian Rahtz]

pah, I hate the web and all its doings.

The story so far.  I have fixed up all the  XSL to generate what seems like proper HTML5 for
the Guidelines, and now I want to check it. How hard can that be? answer; very.

* the validator.nu online service can be driven by a Python script. great.

    - the Python script uses gzip, which needs zlib.  the Python under ubuntu is broke and does not have zlib. gah. #fail, as they say

    - the validator complains that (eg)

	"Guidelines-web/en/html/examples-title.html":95.52982-95.53063: info warning: Text run is not in Unicode Normalization Form C.

	sorry?? is this some new fascism I was unaware of? only NFC form permitted now?

* the RELAXNG schemas at http://syntax.whattf.org/ are unuseable with normal validators, cos they just report

	Guidelines-web/en/html/ref-p.html:113:280: error: no datatype library for URI 'http://whattf.org/datatype-draft'
	required:
		data http://whattf.org/datatype-draft^iri-ref
		data http://www.w3.org/2001/XMLSchema-datatypes^string

You know, I don't have time or energy to recompile Python on the Jenkins server, and ask Martin to do the same.
No, I also don't  have time or energy to compile a Java datatype library and work out how to use that with rnv.  

I feel depressed. I don't want to produce clunky XHTML  that no-one  takes any notice of. I don't want to produce
"any old" HTML which is just guesswork as to whether browsers do the right thing. I don't want to produce HTML5
that seems like something from the Branch Davidians.

I'll make mince pies for dinner instead. at least Lou will benefit.
--
Sebastian Rahtz      
Director (Research Support) of Academic IT Services 
University of Oxford IT Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431