> Personally, I think that it would be an extreme position for the
> TEI to take, to say that that the numeric datatypes in W3C schema
> are inadequate, and it needs to invent its own.

I don't see why this would be "extreme" at all. While W3C may have
done an excellent job in establishing a set of useful generic
datatypes, they may well have left some areas wanting. It is
precisely TEI's role to issue Guidelines about how certain kinds of
information should be expressed. When W3C offers a useful mechanism,
I think we should say "use this W3C mechanism". Where W3C has failed
to offer a useful mechanism, I think we should say "use this method".


> Recommending how to use the W3C datatypes is what I thought we were
> about, and wrapping those recommendations up in TEI syntactic
> sugar.

Becoming a megaphoned cheerleader for W3C datatypes is not at all
what I thought we were about. I thought we were about issuing
Guidelines with recommendations on *how* to encode that which
scholars wish to encode. That may or may not result in a
recommendation to use a W3C datatype in a particular circumstance.


> The advantages of using the W3C datatypes seem to be to outweigh
> the disadvantages of typing 30000000000000.

You may be right, but I am not convinced. I am not at all sure a
project busy encoding papers on astrophysics would agree. It seems to
me the advantages and disadvantages depend a whole lot on a variety
of factors over which you & I have little influence, let alone
control. Who enters the data? How is it validated and proofread? What
processor makes use of it? Does that processor know anything about
W3C datatypes, or is it a home-grown processor that knows about
scientific notation?

While I'm all in favor of using a W3C datatype where they meet our
needs, I do not think they should be a straight-jacket preventing us
from meeting our needs.

I, for one, have never really had to encode a regularization of a
number greater than four digits long. (I've only done it when testing
systems.) I'd like to hear from those who do need to encode such
numbers. How important is it to avoid typing those extra zeroes? How
important is it to avoid them when proofreading? If only the latter
is a problem, a routine that reads in decimal values and writes 'em
out in scientific notation may be sufficient to solve your problems.
(Allowing the datatype to remain a pure W3C datatype, which does have
some advantages, particularly if you're a developer who wants to
process generic TEI documents.)