[tei-council] Guidelines build temporarily borked

[Martin Holmes]

On 14-07-30 08:47 AM, Sebastian Rahtz wrote:
>
> On 30 Jul 2014, at 16:14, Martin Holmes <mholmes at uvic.ca> wrote:
>>
>> The objective was to give a standardized replacement strategy for
>> people moving from @type (where we did declare suggested values)
>> and @rend.
>
> Yes, and the problem there was that the old @type was _locally_
> declared, in a way which we want to get rid of, rather than being a
> member of att. typed, precisely because it had a valList. We are
> reaping our own whirlwind here.

Or, considered another way, we have a useful use-case with which we can 
work out the problems we set for ourselves.

>> I think if there is any value in having the Stylesheets render
>> lists in predictable ways, then there have to be predictable values
>> for @rend.
>
> sure. you’re in TEI Simple territory now! but doing it just for
> <list>, in the base TEI, seems like asking for trouble.

I don't think there's much point in our setting up a new capability 
(overriding a class definition at the element level) if we're unable to 
use it ourselves in our own source.

>> We don't have to declare them in a valList, but if we don't
>> override @rend, we have nowhere to suggest them other than in
>> remarks, which is not ideal.
>>
> you can do it in the prose….

That may be what we end up doing because of time constraints, if the 
release date is soon. But ultimately, we want to be able to do this, surely?

>>> sorting out the underlying processing code to support what you’ve
>>> (which was entirely technically correct) will be hard; I at least
>>> wouldn’t look at it in time for next release, and I suspect it
>>> would be a big ask for someone else to grok odd2dtd in a hurry
>>
>> It depends on the release date, really, and we don't know that yet.
>> The only real requirement is that the rendering for the Guidelines
>> (HTML5, PDF and ePub) be working; all the other outputs can wait a
>> while.
>
> i assume the basic rendering is OK now? that use of @type has been
> supported for quite a whole now.

It seems to be, but I'm a little puzzled by that because I've switched 
all uses of @type to @rend in the P5 spec and chapter sources. The 
Guidelines stylesheets must be capable of handling @rend already (or 
you've fixed them, but the build date suggests otherwise). I'll have to 
look in detail at each of the new values and see if it's rendering 
correctly.

If you look at the header chapter, for instance, the first section has:

<list rend="numbered">

and it seems to be correctly rendering as a numbered list:

<http://teijenkins.hcmc.uvic.ca/job/TEIP5/lastSuccessfulBuild/artifact/release/doc/tei-p5-doc/en/html/HD.html>

> the question is whether odd2dtd can supported your semi-open valList
> change in the source. nothing to do with rendering.

So the suggestion is that odd2dtd is not (yet) expecting a global 
attribute to be overridden at the element level? If that's the case, 
then any ordinary user generating a dtd would presumably hit the same 
problem when trying to do the same thing? And I'd suggest that 
overriding the values of @rend at the element level is precisely the 
sort of thing ordinary users would do, surely?

Mind you, hopefully fewer and fewer ordinary users will be using dtds; 
that may be why we haven't come across this up to now.

Cheers,
Martin

  -- Sebastian
> Rahtz Director (Research) of Academic IT University of Oxford IT
> Services 13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431
>
> Não sou nada. Nunca serei nada. Não posso querer ser nada. À parte
> isso, tenho em mim todos os sonhos do mundo.
>