[tei-council] age= of <person> & <personGrp>

[Syd Bauman]

> > This attribute is not mentioned anywhere in chapter 17 Names,
> > Dates, People, and Places. It is demonstrated (without *any*
> > prose discussion) once in 20.2.2 The Participants Description:
> >   <person sex="2" age="mid">
> > No useful documentation is provided for it in the <person> or
> > <personGrp> tagdoc, and there is one example in the <person>
> > tagdoc:
> >   <person sex="2" age="42">
> > BTW, the same person is being described in the two examples.
> >   
> > The usage in the <person> tagdoc seems to me to be in error.

> The tagdoc is in error, and I have now corrected it to match the
> prose.

Excellent, thanks.


> > The usage in 20.2.2 is what I expected: an alphanumeric string
> > from a project-provided closed set of values that indicate age
> > ranges of interest to the project's research. I can't understand
> > why this attribute is declared as data.word instead of
> > data.enumerated.
> >   
> probably because the enumeration hasnt been formulated yet

This doesn't make much sense to me. We have agreed that
'data.enumerated' means that there should be a documented list of
values, whether the TEI can provide a closed list, only an open list
of suggestions, or has no suggestions at all. So the fact that we
haven't come up with an enumeration yet should not stop us from using
data.enumerated. In such a case its purpose is to signal to a TEI
customizer that she should create a list. There are *lots* of
attributes for which no enumeration has yet been formulated, or ever
will be, that are data.enumerated.


> > But either way, the documentation should be much clearer.
> > Although I don't feel I'm the most qualified to do this, I will
> > be happy to take a first crack at it soon. (But if someone else
> > wants the task, speak up!)
> >   
> 
> I suggest you propose a suggested val|List as a sourceforge feature
> request.

I am not sure a suggested value list is a good idea for this
attribute. Each sociological research group is going to have its own
set of age groups to which its research applies. Heck, even in little
Rhode Island the medico-legal lawyers will consider 'adult' to start
at age 16, the politicians consider 'adult' to start at age 18, and
the liquor stores consider 'adult' to start at age 21!

But I do think this may be a good opportunity to demonstrate the
utility of generating a value list, and how to do so.


> One could also argue that the attribute should be removed or
> renamed -- its original (P4) purpose has been rendered unnecessary
> by the availability of the more pfrecise <event> elements within
> <person> which is probably why it's not discussed in that chapter.

Indeed, age= is a bad name, that most people interpret as indicating
a number of years. ageGroup= or ageBracket= or ageRef= or
demographicAge= or ageDemographicallySpeaking= or ...