i said:
>   so it appears italics might be harder than we think.
>   even two passes, or three!, doesn't assure success.

carlo said:
>   Something that DP discovered in 2004, and has led to the 2005
>   reorganization making proofreading and markup separate tasks.

um, perhaps i wasn't clear, or maybe carlo just missed the point,
but i was saying that, even when we limit the focus to formatting,
heck even _just_ italics, two or three passes might not be enough.

now, if he needs me to be even more blunt, i'll say for carlo that
even the two rounds of formatting at d.p. _might_ not be enough
to assure a questioning observer that all of the italics were found.

(and in case anyone might want to include the proofing rounds,
you should know that proofers are instructed _not_ to do _any_
marking of italics, an instruction which is enforced quite strictly;
moreover, many of the formatters _remove_ formatting that was
done by proofers; there's even a button to erase it automatically.)

i don't believe anyone has checked d.p. output thoroughly enough
to make a judgment as to the completeness of their italics markup.

(and here i'm not talking about the so-called "semantic" markup of
italics which the so-called "best practices" document over there is
now trying to install.  i mean _missing_ the markup, _completely_.)

so, no, carlo, your observation here has no relevance to my point...

***

i understand that jim has informed the list which scan-set he used
for his "huck".  that's good information.  but i don't intend to use it.
my intent wasn't to criticize his performance, or in any way judge it.
i don't think he did worse than any regular person, certainly not me.
(i haven't even done a systematic search for italics myself, and won't.
because i know i'm not too good at it; learned that a long time back.)

the point is, though, that it doesn't do any good for this list to know
the scan-set he used; that information needs to be in the p.g. e-text.

and even there, the information isn't _that_ useful, primarily because
it's extremely difficult to proof a rewrapped text against its scan-set.

in other words, it's not enough to just tell us the scan-set you used.

unless you give the end-users a version of your text which they can
quickly and easily compare to the scan-set, they'll gravitate toward
the entities that _do_ give them the basic capability of confirmation.

but once again, i'm getting slightly ahead of myself.  so, until later...

-bowerbird

p.s.  robert, are you getting all your markup questions answered?
because this probably isn't the best place to be asking about what
needs to be done to make stuff work in the world of .epub/.mobi.
i suggest you instead follow the #eprdctn twitter hashtag, which is
used by people who are actually marking up e-books for a living...