Re: Attribute Architecture -- new type?

On Thu, Aug 14, 2003 at 02:41:34PM +0100, Robert Sanderson wrote:
> * Format/Structure needs 'string' and 'word' term descriptors.

Agreed.

> * All of the Expansion/Interpretations that talk about String or Word are 
>   deprecatable as we only need 'right truncate term' 'left truncate term' 
>   etc.  Any mention of 'string' or 'word' should be replaced by 'term'.

I think I agree - just was not sure what you had intended for truncate
at word boundary and truncate at character boundary. I think truncate
at word boundary only makes sense for 'string' and not 'word' format/structure
values. So 'word' in 'truncate and word boundary' stays 'word' and is not
changed to 'term'. Otherwise I agree.

> * 'any', 'all' and 'adjacent' are comparisons that apply to multiple 
>   terms.

I agree.

> * 'single' and 'multiple' are the structure of the Term, and hence need a 
>   new attribute?
> 
> 'Number of Terms' attribute might be something like:
> 
> (integer)    Exactly that number of terms
> 'null'       No terms
> 'multiple'   Unknown number of terms, 2 or more
> 'unknown'    Unknown number of terms, 0 or more

I understand what you are saying, I just would like some examples to
be convinced its necessary. Surely the server is going to parse the
supplied text to work out what terms are in it? Would knowing the
number of terms change how the parser did its job? If parsing words
from a string, how would the client know the exact parsing rules
the server is going to use? (eg: book-case - how many terms?)

So I agree there are zero to many terms that may get extracted from
the supplied InternationalString/GeneralString/etc, but I am not sure
what use there is for the client to supply how many there are.
And what if the client gets it wrong? But I am happy to be convinced.

Alan

Received on Friday, 15 August 2003 05:14:12 UTC