[tei-council] <quotation>

[Sebastian Rahtz]

On 11 Dec 2012, at 10:11, Lou Burnard <lou.burnard at retired.ox.ac.uk>
 wrote:
>> as you'll probably have seen on TEI-L,
> 
> (If I didn't know you better, I'd find the tone of your message there a 
> little offensive, btw)

Yup. Apologies. Error of tone in my message.


>> I'd like to ask permission to remove that <defaultVal>, for two reasons
>> 
>>   a) we can't implement default values in RELAXNG schemas
> 
> Tail wagging dog?

well, yes, there's an element of that I confess. but  we have gone so far
in the last few years to make the Guidelines outputs closer to being
implementable (datatyping, schematron etc), that I think <defaultVal>
is looking a bit out of place now. 

>>   b) because <quotation> is optional, the meaning of the "default" is ambiguous
>> 
> 
> It's not ambiguous: it's underspecified. They are not the same thing. 
> "Underspecified" means an implementer can make up their own mind what to 
> do in that case. "Ambiguous" means they cannot because they are given 
> contradictory pieces of information

I'll accept the word may be wrong, but the effect is the same. The creator
of the <text> does not know what will happen in the rendered output if she omits
<quotation>. If she chooses to write <quotation marks="some"> the
she'll get her sandy deserts of course.

> 
>> I note that the Guidelines examples seem universally to
>> NOT retain quotation marks.
> 
> "universally" meaning "nearly all of the time" I assume since your TEI-L 
> message points to a case where they *are* retained.

I did take the liberty of removing the one I found in <said>. If people
think that was bad, I'll restore them.

> Anyway, yes, I think 
> it reasonable to assume that <q> etc. normally replace the punctuation 
> marks which triggered the assumption that this stretch of text should be 
> treated as  "quotation" (that's what the definition says anyway).

yes, that seems to be the tone of the discussion, and backed up
by our own examples. In which case the (possibly) implied default of <quotation marks=":all">
is contradictory.

>   On the 
> other hand, some people think (equally reasonably) that source 
> punctuation is interesting and should be retained, so you need some way 
> of being told which doctrine obtains in a given text.


sure. and thats what <quotation> lets you do. I have duly implemented
a check for that in the XSL.


So do you agree that the simplest thing to do is remove that <defaultVal>,
as it arguably contradicts the prose?
--
Sebastian Rahtz      
Director (Research Support) of Academic IT Services 
University of Oxford IT Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431