[tei-council] adding (non-)Struct elements to <date> & <time>

[Lou Burnard]

Syd Bauman wrote:
> SB> Why is <offset> in this list at all? I.e., I think maybe <offset>
> SB> should not be a member of att.datePart.
>
> LB> It's there so that you can mark up things like "ten days after
> LB> Easter".
>
> I'm not sure, but I think maybe you're answering the wrong question.
> I'm not asking why does <offset> exist, but why is it a member of
> att.datePart. I.e., why would you want to use value= or full= (or for
> that matter, even type=) to encode the word "after":
>
>   
Yes, quite. It shouldnt be a member of att.datePart

>    <date value="2006-04-26">
>      <distance value="P10D">ten days</distance>
>      <offset>after</offset>
>      <name type="holiday">Easter</name>
>    </date>
>
>
>   
>> There is an argument for saying we should throw all that clever
>> stuff from time/dateStruct away and use something like TimeML
>> (http://www.timeml.org/site/index.html) instead
>>     
>
> Hmmm... haven't taken a careful look, but my gut instinct is no, we
> would not want to just use TimeML instead, but it is probably a good
> idea to look through it for ideas. E.g., their @mod attribute may be
> useful. 
>
>   
I actually said "something like TimeML"

>
>> If we're not sure why it should be there (and thinking about it
>> again, I'm even less sure than I was), then let's not have it at
>> all.
>>     
>
> Not sure what you mean here:
>
> * If we're not sure why <offset> should be in att.datePart, we should
>   take it out of att.datePart (my point)
>
>   
Sorry, something got missed out from my post (too much editing going on 
in parallel). I wasn't talking about <offset> any more; I was talking 
about the "type" attribute.