> How does TEI handle indentation patterns and whitespace?
TEI does not really directly address this issue at all. It is
indirectly addressed in that indentation is a renditional feature to
be handled by the all-in-one rend= attribute. Projects are left to
themselves to decide how to use the rend= attribute.
At the WWP we indicate levels of poetic indentation in the abstract,
using the rend= attribute.
<l rend="indent(1)">O, tell me, what are life's best joys?</l>
<l rend="indent(2)">Are they not visions that decay,</l>
<l rend="indent(1)">Sweet honey'd poisons, gilded toys,</l>
<l rend="indent(2)">Vain glittering baubles of a day?</l>
-- WWP TR00190 1.105 lines 4679-82, inclusive
> If a stanza has a regular stanzaic form can one capture indentation? Or do
> we encode unicode tabs (is that 	 ?).
I'm not sure what you mean by "regular stanzaic form". If what you
mean is that stanzas have predictable patterns of indentation (e.g.,
lines 1, 3, 5 are indented, lines 2, 4, 6, & 7 are not) then I
suppose you could capture that information on the rend= of <lg>
instead of the rend= of <l>. Or, if you consider this more vital you
could create a new attribute for this information. If you mean that
there is some formula from which indentation can be derived (e.g.,
feet + amount indented = constant) then you could certainly leave the
information out, and have processors calculate the indentation from
the number of feet (as shown on the met= or real=), but this runs
afoul when there are mistakes or ambiguous cases in the source.
I personally like to shy away from encoding whitespace (other than
that which separates words) as data instead of in the markup.
> I'm working in XML so I presume I can ignore named entities and
> just code everything in UTF-8?
Certainly. If your editor will do the right thing you could just type
a tab. (Although again, I prefer indicating indentation in the
markup.) I don't think there is an ISO named entity for horizontal
tab, is there? In which case, after
<!ENTITY htab "	">
all of the following should be equivalent
where "@" indicates an actual tab in your file (i.e., "00001001" in
binary UTF-8 -- a bad idea, as you can't see it and lots of editors
muck with 'em; but if you're hell bent on entering one, in Emacs use
> Can one specify a font and font size or FO characteristics?
You certainly could (e.g., rend="face(Century Schoolbook)
size(10.5)"); alternatively one could specify an abstract identifier
on rend=, and then associate that with particular formatting in a
stylesheet or other process (e.g., rend="poe07").
> Poetry can stretch the idea of separating content and style, as the
> style can be the content in experimental poetry.
Indeed. Shape poems are enough to keep me up at night. :-)
 Note that the hex and decimal values just happen to be equivalent
because the value for tab happens to be less than 10. The same
list for the symbol used to represent a horizontal tab is
which I believe would be "111000101001000010001001" in UTF-8.