[tei-council] @xml:space, whitespace and http://purl.org/tei/bug/3600991

[James Cummings]

Some of the confusion, I think, might revolve around the 
dichotomy (which TEI exposes as false) between a simplistic view 
of an element's content as 'structured' or 'mixed content'.  The 
TEI allows the same element to act as both and <persName> is 
indeed a good example of this. It sometimes just may have 
<forename> and <surname> in it, but other times might also have a 
comma or just text. If it is not mixed content then the tests 
I've done indicate it is usually treated as if it is what John 
calls a 'structured' element.  To do otherwise implies 
schema-aware processing. There are plenty of times I've treated 
<persName> as if it is a 'structured' element... in fact I prefer 
to do that.  I've never used @xml:space to indicate my intention, 
just relied on things following the same assumptions as me.  (I'm 
not saying this is a good idea, just reporting on my usage.)

I think I'd just remove the sentence starting 'If it is done,...'.

-James

On 16/01/13 13:44, Martin Holmes wrote:
> The text reflects my own understanding of the situation (although I'm as
> cautious as Sebastian with regard to whether I do understand it or not).
> One bit caught my attention, though:
>
> "Treating a structured element as a mixed-content one, or vice versa,
> should be done with care. If it is done, the schema should be customized
> to record the fact."
>
> How would one customize the schema to do this? I can see that you could
> use @xml:space="preserve" to specify that an element with element-only
> content has significant whitespace, but that's in the encoding itself,
> not the schema; and I don't see how you can specify the reverse (since
> "default" won't do it). The text also explains how you can use
> <encodingDesc> to document such things. But I'm not sure how you could
> do it in the schema.
>
> Other than that, I think it's great.
>
> Cheers,
> Martin
>
> On 13-01-16 04:30 AM, James Cummings wrote:
>>
>> Hi there,
>>
>> John McCaskey believes we introduced new errors and confusion by
>> our attempt to clarify the use of the @xml:space attribute and so
>> re-opened a ticket about it.  I suggested he rewrite the copy to
>> what he believed would express the truth and Sebastian and I have
>> both looked at it.
>>
>> http://purl.org/tei/bug/3600991
>>
>> Since this is a textual (non-schema affecting) change, could
>> Council look at this ASAP and see if you think it represents the
>> truth of what we believe @xml:space to do?  A lot of this seems
>> to originate from vagueness in the original specification and
>> different tool implementations.
>>
>>
>> -James
>>
>>
>>
>


-- 
Dr James Cummings, James.Cummings at it.ox.ac.uk
Academic IT Services, University of Oxford