Some notes on Chapter 9, Base Tag Set for Verse

Relatively substantive comments:
In the examples with <lg>s, starting with the one from Ginsberg's "My
Alba" in 9.2, the first <lg> is marked type=free, but the second is
type=CDATA.  This doesn't make sense -- it should also be "free".  

type=CDATA occurs in several places: it looks like a formatting error.

In P3 the type of an <lg> was #CURRENT, which was convenient; are you
assuming that applications processing P4 documents are supposed to
take the default value "CDATA" to behave as if it were the last
non-trivial value?  In other words, this example looks to me as if my
applications processing P4 documents (at least those in XML) have to
implement #CURRENT themselves.  It might be easier to put the desired
value explicitly in the attribute in the examples.

#CURRENT isnt legal in XML, so yes, this is something that an
 application has to work out for itself.

The second example in 9.3, from Virgil, in its SGML incarnation was very 
fussy about spacing, with spaces between words but not within words.  The
P4 version has white space after every syllable.  Perhaps something more
like this:
<l>
<seg type="foot">
   <seg type="syll">Ar</seg><seg type="syll">ma</seg>
   <seg type="syll">vi</seg></seg><seg type="foot"><seg type="syll">rum</seg><seg type="syll">que</seg>
   <seg type="syll">ca</seg></seg><seg type="foot"><seg type="syll">no</seg>
   <seg type="syll">Tro</seg></seg><seg type="foot"><seg type="syll">iae</seg>
   <seg type="syll">qui</seg></seg>
<seg type="foot"><seg type="syll">pri</seg><seg type="syll">mus</seg>
   <seg type="syll">ab</seg></seg>
<seg type="foot"><seg type="syll">o</seg><seg type="syll">ris</seg>
</seg>
</l>

It's a whole lot less readable this way, but if you're marking every
syllable, readability is hardly likely to be a goal!

-> I have restored the missing spaces

The next paragraph discusses the division of the hexameter into two
"hemistiches."  The usual term in classical meter is "colon" (and it's
customary to use the Greek plural form "cola"), and it is almost
invariably true that the colon boundary comes within a foot.  (It
might be after the fourth foot, but the normal position is inside the
third foot.)  Hence this is not especially "an interesting case" as
the second sentence of the paragraph says.  The quickest fix is
"typical" instead of "interesting," I suppose.  (This isn't new with
P4, of course.)

-> Thanks for the correction! I can now confess that I only just
scraped through my O-level latin (tho I did have to do Vergil as part
of the Oxford English course!)

About the "enjamb" attribute at the end of 9.3: The conventional
values 'y' and 'n' are fine, but for P5 what about an example
distinguishing "necessary" and "unperiodic" enjambement?

--> &winita;



The last example fragment at the very end of 9.4.3 says the metrical
scheme for the "commiato" is E/S/S/E/S/E/E/S/S/E/E, but the first line
of the text looks like an S because the last three words (e\ una
donna) have been omitted; this is confusing to a reader who doesn't
know the poem.

->restored the missing words

In section 9.6, another prosodic feature that might be worth marking
is elision.

-> &winita;

Much of 9.4 implicitly assumes foot-based meters, as in English,
though it would in fact be possible to encode Greek lyrics in this
scheme.  I have not actually done so; at Perseus, we generally go to
the opposite extreme, using only the name of the metrical family in
the 'met' attribute.  For example, our Catullus 1 is marked 
    <div1 type=poem met=Phalacean>.  
In P4, I'd recommend indicating that this is one possible use for this
element.  For the future, some more complicated examples would
demonstrate whether the <metDecl> and 'met' attribute are really
practical for elaborate metrical schemes.  It's clear how to use this
for the simple stanza forms of most English verse (and most Latin), or
for stichic verse in any language.  I'm thinking about the lyrics of
Greek drama, or of poets like Pindar and Bacchylides (Sappho and
Alcaeus are easy), or in English perhaps Cowley.

--> &winita

Typographical, grammatical, stylistic:
Chapters and sections are referred to throughout the HTML version in
the form

"... described in section 6.11.1 Core Tags for Verse include...."   

It might be more graceful to put commas around the chapter name, as

"... described in section 6.11.1, Core Tags for Verse, include...."  

The PDF version has a comma between the number and the name but none
after the name, which is incorrect.

Yes, something needs doing!


Spaces needed: last bullet in third paragraph of introduction, in
"and9.5 Rhyme" between "and" and "9.5"; fourth paragraph of
introduction, "section3.3 Invocation" between "section" and "3.3"; end
of introduction, "chapters6" between "chapters" and "6".  One space
too many in 9.6, second paragraph, "Feature Structures ,".

I note that the header comment in the DTD fragments refers to "TEI P3,
1994"; I'm sure I'm not the first one to observe this.
-> Actually...