[tei-council] Fwd: grouping elements under <choice>

[Lou Burnard]

As Peter says, we shouldn't start the fight again. Or not until there 
are some new arguments to muster anyway. But I can't help myself...

<choice> is meant to be used for VERY SIMPLE CASES where you want to 
indicate that there is More Than One Way of Doing It. Each of the 
component children of a <choice> should represent one way of doing it.
The types of such components are deliberately restricted to
-- various kinds of editorial (non)intervention
-- generic <seg>
-- <choice> itself
This does make it possible to indicate alternative choices within 
choices -- but not aggregations of choices within choices, which is 
Piotr's concern.

I don't understand the grounds of Piotr's objection to the 
"supersegmentation" solution though. In one case, the sequence "2." is 
to be regarded as lexical; in the other it isn't. So why not
<choice>
<seg type="lexical">2.</seg>
<seg><seg type="lexical">2</seg>
      <seg type="punct">.</seg>
</seg>
</choice>
?
(obviously /<seg type="lexical">/ here is syntactic salt for <w> and 
/<seg type="punct">/ for <c> )



Peter Boot wrote:
> I would think it is an important issue. But it also has wide-reaching 
> implications. Piotr asks for a way to represent multiple analyses of his 
> text. But a next case would obviously be to represent multiple 
> transcriptions, in case of doubt. And then the next option would to 
> store a text and its translation under choice. etc. But where does it end?
> 
> If we're going to do something about this I would think it needs a small 
> working group to sort the issue out.
> 
> (And if you already had a fight over this, why start it anew?)
> 
> Peter
> 
> Laurent Romary schreef:
>> Hallo Council,
>> This issue is something that has bothered me since quite a while with  
>> somehow having the same kind of proposal in mind. Is this an issue on  
>> which the council would have some strong feelings (for or against)?  
>> Could/should it be put on our working agenda?
>> Cheers,
>> Laurent
>>
>> Début du message réexpédié :
>>
>>> De : Piotr Bański <bansp at O2.PL>
>>> Date : 15 décembre 2008 23:28:10 GMT+01:00
>>> À : TEI-L at listserv.brown.edu
>>> Objet : grouping elements under <choice>
>>> Répondre à : Piotr Bański <bansp at O2.PL>
>>>
>>> It seems to me that the <choice> element is only able to provide
>>> alternatives between atoms, and that it lacks a bracket-like child,
>>> something like <group> in RNG and a few other places.
>>>
>>> I'd like to mark up a choice between two alternative segmentations:
>>>
>>> either:   <seg/>
>>> or:       <seg/><seg/>
>>>
>>> and there appears to be no easy way of expressing this.
>>>
>>> I'd like to have:
>>>
>>> <choice>
>>>   <seg/>
>>>   <group>
>>>      <seg/>
>>>      <seg/>
>>>   </group>
>>> </choice>
>>>
>>> but I can't. I can use a kludge and wrap the two segments into another
>>> <seg/> rather than <group/>, but that would be claiming that the two
>>> segments form a unit -- that, however, is not the case here.
>>>
>>> Consider a sequence such as "2.". In Polish orthography, it can mean
>>> either "the 2nd" (<seg>2.</seg>) or it can be the number "2" standing
>>> right before the end of the sentence (<seg>2</seg><seg>.</seg>). I  
>>> don't
>>> want to claim that in the latter case, I am dealing with a single
>>> "super-segment" that happens to contain <seg>2</seg> and <seg>.</ 
>>> seg> --
>>> they are not a super-segment, they are a sequence of two segments.
>>>
>>> There are other examples that involve morphemes, and it seems to me  
>>> that
>>> a similar need must pop up in the case of other element sequences as
>>> well -- it feels rather generic.
>>>
>>> Is there any clean way to capture such alternatives, or may this  
>>> perhaps
>>> qualify as a feature request?
>>>
>>> Thanks,
>>>
>>>  Piotr
>> _______________________________________________
>> tei-council mailing list
>> tei-council at lists.village.Virginia.EDU
>> http://lists.village.Virginia.EDU/mailman/listinfo/tei-council
>>
>>
> 
> _______________________________________________
> tei-council mailing list
> tei-council at lists.village.Virginia.EDU
> http://lists.village.Virginia.EDU/mailman/listinfo/tei-council