[tei-council] Re: roma, documentation, i18n

[Lou Burnard]

It's also a philosophical point. ODD is "one" document does it all. In a 
semi-normative document like P5 you need to have one canonical 
translation for each language, and the logical place to store that 
canonical translation is in the canon. I see the language module files 
which Sebastian is talking about as being exactly analogous to other 
non-canonical schema modifications.

Theres a whole lot of quality control issues here too, of course: at 
what point and how does who decide that translation x is canonical, and 
can therefore be merged into the source?

I also think it's important for a useful to be able to see other 
translations. If, for example, I were translating into Italian, I would 
probably find a pre-existing French translation helpful.

Lou


 Sebastian Rahtz wrote:

>Christian Wittern wrote:
>
>  
>
>>What would be the advantage of pulling this into the main P5 source?
>>I see only added maintainance cost with no real benefit.  You could as
>>easily pull them out and merge them into the tree on runtime.
>> 
>>
>>    
>>
>perhaps its matter of semantics; if the main ODD processing
>software expects to find the various language-specific <desc>s
>in site in the main document, then it does not much matter whether
>the merge is at runtime or at source. I just want to avoid a "normal"
>ODD processor from having to know about the language modules.
>
>  
>
>> 
>>
>>    
>>
>>>dont forget attributes, and vallists, by the way
>>>   
>>>
>>>      
>>>
>>But attributes should be covered within the elementSpec, shouldnt
>>they.
>>
>> 
>>
>>    
>>
>and the classSpecs.
>
>
>  
>