At 22:43 2004-07-19 -0400, Michael Priestley wrote:
>Thanks for the clarification. Sorry if I jumped too soon on a non-issue.
>
>In last week's discussion, wasn't the thought that we should try to provide
>whatever the widest possible tag set would be, as a base?
There were several "thoughts" at last week's telcon that went beyond
our previous conversations about table model. I didn't think any
particular idea that came up last week necessarily trumped all our
previous discussion on the topic.
In particular, I would not vote for using the widest tag set as a base.
For example, I see no reason to have both spanspec and namestart/end
(which complicates the model and allows for self-conflicting markup)
when spanspec's in legacy data can programatically be converted into
namestart/end.
A lot of effort went in to developing the OASIS Exchange model, mostly
because the CALS model was ambiguous (sematically) and underspecified.
I still favor standardizing on the Exchange model as a good base to use.
paul
> Admittedly this
>puts more of a strain on the base transforms, but it would then allow the
>widest possible range of variations through specialization.
>
>Michael Priestley
