[tei-council] some attribute issues

James Cummings James.Cummings at oucs.ox.ac.uk
Wed Oct 17 05:39:46 EDT 2007


Syd Bauman wrote:
>>> * I hope I don't sound like a broken record, but the version=
>>>   attribute of <TEI> is still declared as data.decimal. (Besides not
>>>   doing cool things like pointing to the ODD, which IIRC James wants
>>>   and is what derailed the conversation last time, this doesn't even
>>>   permit "1.0.1", or "5.0.1" or "P5 1.0".)
>> What should it be then?

For the record, I wanted to point to the ODD I used from encodingDesc 
which, I suppose, I can do by doing something like ab/ptr/@target inside 
it.  encodingDesc seems more appropriate for that kind of information than 
TEI/@version.

> Well, unless we can agree on a version numbering scheme in *very*
> short order (for which I don't have high hopes, as we have been
> unable to do so for years), I think it has to be 
>   <datatype maxOccurs="unbounded">
>     <rng:ref name="data.word"/>
>   </datatype>
> no?
> 
> Personally, I'd much prefer we dictate what should be used as the
> version number. E.g. "P5" followed by ":" followed by either 
> a) the Sourceforge release number, e.g. "0.4.1",
> b) the Sourceforge revision number, e.g. "r3682", or
> c) the date of the revision, e.g. "2007-05-23".
> Or some such system.

While I wouldn't complain about data.word, I would prefer something which 
allowed a or b.  I.e. those using SVN can put version="P5:r3682" those 
using public announced releases can put version="P5:1.0.23".  However, 
wouldn't it be possible in generating the schemas to force the value of 
this attribute to be the TEI version which the schemas were created against?

> Back to ST for a bit before bed.

I will be looking at this today sometime, promise.

-James
-- 
Dr James Cummings, Oxford Text Archive, University of Oxford
James dot Cummings at oucs dot ox dot ac dot uk


More information about the tei-council mailing list