David Megginson wrote:
>
> As I recall, the idea of adding the 'type' attribute to %a.global; has
> come up more than once on this list as a passing suggestion. Is it,
> perhaps, time actually to implement this change?
I think I agree with Dominic Dunlop that caution be exercised here. It's
easy to add a global attribute, but difficult to control the semantics
(as satisfying different user communities), and impossible for SGML to
validate the semantics in any case. To wit:
1) If type="" becomes a global, then TEI or the different user communities
must specify for each element some recommended taxonomy of types, else
the encoding is meaningless. I can't imagine anything other than CDATA
for type="".
2) Specifying a taxonomy of course means adopting a certain analytical
perspective (or several), and since these differ within user communities,
it may be difficult to achieve consensus on anything meaningful. If
consensus is not desirable, then we place increased burden upon the
declarations features of TEI, for which there is little software support,
as far as I know.
3) In many cases, it will also be necessary to revisit the current
usage of n="" in relation to the proposed type="". Some groups have used
n="" because type="" was not readily available -- and perhaps legitimately
in the sense that names of things as values for n="" can plausibly
represent a type hierarchy. The relationship between the two would
need to be explored and clarified for many elements, I would imagine.
4) My private concern is that the encoding practices available in TEI
are perhaps already too under-specified in some domains (meaning:
there are usually many ways to encode the same information), making it
even more difficult, potentially, to implement generic "TEI text
processing software." If so, addition of global type="" might
exacerbate the problem.
Robin
-------------------------------------------------------------------------
Robin Cover Email: [log in to unmask] ("uta-ef-el-el")
6634 Sarah Drive
Dallas, TX 75236 USA In case of link failure, use:
Tel: (1 214) 296-1783 (h) [log in to unmask]
Tel: (1 214) 709-3346 (w)
FAX: (1 214) 709-3380 SGML Page: http://www.sil.org/sgml/sgml.html
========================================================================
|
