[tei-council] Re: on spec grp 2, datatypes normalized

[Syd Bauman]

> This is indeed the case, and reflects a conscious decision to use
> attribute values here as elsewhere to represent a normalized value,
> which might indeed be represented in a more nuanced way within the
> content of the element or by linking elsewhere.

While I agree with the general principle, "13:30" is not more nuanced
than "13:30:00"; but it is different, and needs to be representable.


> I don't get it. If you want to give a value precise to only a year,
> month etc. you can do so using the approved designation, surely?
> "1900" doesnt mean any particular time in 1900, it just means 1900.
> Similarly, 1900-12 means some time in December 1900. And so on.

Exactly! Using W3C datatypes you can give a date/time value precise
to the
* year
* month
* day
* second
but *not* to the hour or minute. This is what requires repair.


> Surely, the point of using this format is precisely that as many
> processors as possible will do the right thing with them.

Yes, but not at the expense of not being able to represent humanities
data. Besides, keep in mind that (at least for RelaxNG) all a datatype
does is, given a certain string, say "yes, this is" or "no, this is
not" a valid instance of me. It doesn't return a normalized or
regularized or canonical value or anything like that.


> Can you give some examples of processors which do not behave
> correctly when handling ISO 8601 format values? And is there any
> reason to believe they would make a better fist of handling an
> arbitrary TEI-invented format instead?

The format I'm recommending is *not* arbitrary and is *not*
TEI-invented; it *is* ISO 8601.


> Similar considerations apply to the other two cases you mention. In
> the case of duration and sex, we are talking about an encoded value
> for an attribute, and it makes much better sense to use an ISO- or
> W3C- recommended encoded value than to make one up.

[On duration; ignoring sex, where it seems we're agreed on ISO's 0, 1,
2, 9:]
Again, it's only syntax, so I'm not claiming it is very important, but
I'd like to set the record straight. I did not make up the syntax I am
suggesting. It comes from NIST which takes most of it from SI. I
daresay the Bureau International des Poids et Mesures and even the
(US) National Institute of Standards and Technology has been doing
this standardization of units stuff for a lot longer than either ISO
or W3C, and they do a significantly better job.