> Wouldn't you use <hyphenation eol="some"/> in the header if you
> decide to remove the ones you're "commited to"?
Well, yes, I probably would. But how would that help the poor program
trying to reconsitute words in my document? (The legal value list of
eol= really doesn't do the job.)
> As to the tokenization, @break does this for you (more or less)
Well, yes: less.
> evil-<lb break="yes"/>doers -> tokenize as evil- doers (or "evil
> doers" if you decided to suppress the hyphen too)
OK. Still leaves the poor program trying to figure out whether to
supress the hyphen or not, but this is not a use-case I imagine would
> evil-<lb break="no"/>doers -> tokenize as evil-doers
Why as "evil-doers" and not as "evildoers", which is what hap-
pens with "hap-<lb break='no'/>pens".
> evil-<lb break="maybe"/>doers -> if you don't want to take up a
> position at all.
Right. (And since I'm wishy-washy, *I* might use this. :-) But this
still leaves the program trying to reconstitute words in a lurch, it
just does so on purpose.
> Back in the main thread, however, yes I totally agree that the
> so-called Popham Proposal (if he's really responsible for it) is a
> Very Bad Idea. I suspect it may have been a consequence of a data
> model which didn't permit milestone elements outside divs.
> Note, however, that just saying "we always encode <note>s inline,
> even when their content spans a page" is OK if your objective is
> "find out what page I begin on", but less so if you want to do
> something like "count all the words on this page, whether within
> notes or not"
Agreed, at least if you don't indicate the page break. (I'll post
about that soon.