Francois Lachance wrote:
> The witty exchange between Lou and Sebastian prompts me to ask Lou, now
> back from holidaying, if he would kindly provide a neat little
> disquisition on the element classes "inter" and "divn" [my copy of the P3
> Guidelines being blank at page 780 -- I know I should consult P4 online].

You could do that (http://www.tei-c.org/P4X/ref-DIVN.html), or even
better you could buy a copy of the print version and do your bit to keep
the TEI financially afloat...

The thing to remember is that there are two different kinds of class in
pre-P5 versions of the TEI: model classes (e.g. inter, divbot) which are
used to construct content models; and attribute classes (e.g. metrical,
divn) which are not used in the content models, but to indicate that
members of the class share a set of attributes. (Then there are a few
classes which are both, but that's another story). This is detailed in
excruciating detail in chapter 3 (http://www.tei-c.org/P4X/ST.html#STEC)

>
> The index to P3 does indicate that there are three div-related classes
> (in alphabetical order):
>
> divbot
> divn
> divtop
>
> Yup, I checked P5 -- no "div" class (Yes, I know that the "divn" class
> references <div> elements )

I dont understand what you mean. <div> is one of the elements which is a
member of the divn class, because it carries the attributes defined for
said class.

>
> I am being a little stickler for a good reason. Would knowledge of divbot
> and divtop classes help some encoders trying to place that extra <div>
> after a <p>?

Nothing can help such people, I fear, if you are talking about those who
want to go down and then up again inside a div.

>
> At present the Guidelines present the element classes atomistically (as a
> descriptive alphabetical listing). Does anyone have a nice chart depicting
> the relations between the various classes?

Not that I know of, though it would certainly be an interesting exercise
to try to ake one.

If I read P4, correctly <ab> is
> a member of the "chunk" and "typed" classes. Lou perhaps can enlighten me
> as to the differences between ~levels~ and ~classes~.

"level" is a specific and rather restricted kind of class. We tend to
talk about phrase-level elements (which can appear inside
component-level objects),  and inter-level ones (which can appear either
within or between component level objects), and about component-level
objects which can appear inside a text body. There are many more proper
model classes, and they have a wider semantics than simply hierarchic
position.


  I am
> partially perplexed by the following statement. I understand that <salute>
> and <ab type="salute"> are equivalent but <div type="salute"> would be
> possible, no, in a <div> within <div> use case? ...
>

Well, only partially perplexed is not so bad. I apologise for making an
obscure joke in public! Anyway, to answer your question: yes you could
indeed have <div>s all the way down if you wanted. But the key point
about the divtop elements is that they are constrained to appear either
at the beginning or the end of the div. Look at the content model for
div, and you will see that (simplifying somewhat) it allows an optional
sequence of divtop elements, followed by an optional sequence of
component level elements, followed by a recursive div. As <salute> (and
hence <ab type="salute">) is a divtop element, you shouldn't put it
anywhere a <div> (or for that matter an <ab>) is legal.

The TEI class system will all be quite different, and possibly easier to
understand, in P5 so I wouldnt worry about it too much for now...


>
>>Certainly not. They would, if anything, be <ab type="salute"> elements,
>>since they are all inter level, not div level. I'm surprised at you!
>>
>>L
>
>
> Fascinating game of translate this into that.
>
>
> -- Francois Lachance, Scholar-at-large
> http://www.chass.utoronto.ca/~lachance
>
>
>
> Wondering if...
>
>         mnemonic is to analytic
>                  as
>         mimetic  is to synthetic
>
>