[tei-council] @source and @version for versioning an ODD

[Kevin Hawkins]

Thanks to everyone for being so responsive while I was touring Paris. 
Responses to parts of a few messages below:

On 11/6/11 7:55 AM, Sebastian Rahtz wrote:
>> Please revise the wiki accordingly: the point of creating this page was
>> that I am trying to take the first steps towards documenting this
>> capability, which, as I originally wrote, does not seem to be documented
>> anywhere.
>
>
> I am a bit puzzled by this. I don't say its well done, but the guts of it
> are there in http://www.tei-c.org/release/doc/tei-p5-doc/en/html/TD.html#TDbuild
> where they oughta be. I don't think the wiki article takes the user further forward
> than the existing Guidelines.

Ah!  This is what I was looking for!  Since I wrote to Council about 
this on 20 October, I had been looking for information on versioning an 
ODD in section 23.2; I'm afraid it never occurred to me to look in 
chapter 22!  And as I whined then, I wasn't able to trace the history of 
the change in SourceForge.

Perhaps we can get a cross-reference to #TDbuild from section 23.2?

On 11/6/11 8:05 AM, Sebastian Rahtz wrote:
 > There is no @version attribute on<schemaSpec>  is there? its 
definitely @source.

You're right.  I took what Syd wrote to me exactly without verifying. 
I've corrected in the wiki.

On 11/6/11 8:25 AM, Lou Burnard wrote:
 > Let me try to clarify the confusion that seems to underly this thread.
 >
 > Firstly on @version versus @source : these attributes serve different
 > though related purposes, and are provided by two different classes.
 >
 > Members of att.translatable (as the name suggests) are low level text
 > containing elements within a spec which contain text that has been or
 > might be translated. The purpose of the attribute is to specify the
 > version of P5 which was current when that translation was effectuated.
 >
 > Members of att.readFrom however are specification elements like
 > <schemaSpec>  or<elementSpec>. The purpose of the @source attribute is
 > to specify the version of P5 in which the specification is to be found
 > (or, more exactly, in the case of schemaSpec, from which by default its
 > child specification elements are to be found.
 >
 > Secondly, the *value* of both attributes is declared as a specific
 > datatype, conforming to some standard or other (cant remember which)
 > used to indicate a version number of three digits.

See, the datatype of att.readFrom (@source) is xsd:anyURI whereas the 
datatype of att.translateable (@version) is data.word.  As I noted in 
the wiki, I wonder if this is a bug.

Well, I'm about to head out for a bit more touring.  Till tomorrow,

Kevin