Em dashes not treated as punctuation

[Mark Dewey]

I've noticed that em dashes aren't treated as punctuation, but rather as 
letters (and thus part of the words).

I can see some reasons why this may have been overlooked, though:
1. Em dashes require special actions to type (and in LilyPond they would 
probably be encoded as UTF-8 characters, though they are also in the extended 
ASCII set).
2. In normal writing (outside of LilyPond) there are not spaces between em 
dashes and the words surrounding them. (so, typing word1—word2 in a lyrics 
context would treat it as a single word)
3. Not everyone is consciously aware of em dashes (although they are very 
commonly used anyway).

Anyway, this is the character I mean: —
On a PC, create it by holding alt and typing 0151.

I don't know if they're really called em dashes, but I saw some website call 
them that once.

Anyway, ideally, in a music notation program, there would be no space between 
the em dash and the words around it.  Sometimes, however, this may not be 
practical, but in those cases it seems that unless the em dash is separating 
off a segment ending the sentence, there should be equal space between the word 
behind it and the word in front of it.  If there is only one em dash in a 
sentence, (or if it is separating off a segment to end the sentence) it seems 
it should at least be placed tightly next to the left word, with the space 
(this part alone is easy enough to do, if you don't mind the space after it).

Anyway, I haven't seen a way to make this how I've generally seen it done in 
printed sheet music, in LilyPond, so far, but it would definitely make for a 
more professional look.

Here are some example sentences with how it seems the em dashes should be 
formatted:
1. I went to the store—so did she, at that.
(If the above couldn't be so tight, the ideal would be to have the space after 
it: I went to the store— so did she, at that)
2. We were looking for a blue car—Susy wanted a black car—for the show.
(If the above couldn't be so tight, it seems the ideal would be to have equal 
space around the em dashes: We were looking for a blue car — Susy wanted a 
black car — for the show)

Also, no space between the dash and the left word if the right word is on 
another line:
i.e. first word—
next word

Here are some examples of how it seems it shouldn't look:
I went to the store —so did she, at that.
I went to the store — so did she, at that.
We were looking for a blue car— Susy wanted a black car— for the show.
We were looking for a blue car —Susy wanted a black car —for the show.
We were looking for a blue car —Susy wanted a black car— for the show.
We were looking for a blue car— Susy wanted a black car —for the show.
etc.

Whatever the case, it looks much better (much more professional, and much more 
like it looks in literature) having them as tight as possible, whenever 
possible.

Right now, the only fix I know of seems to be to always add some spacing (which 
is very difficult to make equal, as it must be done differently every time, and 
this can shift when something else about the music does), unless there is only 
one separating em dash.
 But it should be noted that these are fixes, and probably not what should be 
the end solution (especially for how often these things show up).  Anyway, I 
don't think Finale does this either (or any other computer notation program), 
but it seems calculable enough to integrate without problems.

Maybe you want to talk to an expert on these dashes before taking my views on 
what it seems they should and shouldn't look like to heart, though, but I'm 
sure they'll at least agree that it is ideal to have them as tight as possible 
on both sides.  My views primarily come from what I've seen done in musical 
printings (not from computer programs for notating music), and in printed 
literature.

I don't know if you could (or would want to) actually make it so the character 
itself was treated as punctuation.  Maybe it would be better to have some 
simple code to input one between two words (that way you perhaps wouldn't have 
to deal with UTF-8 stuff so much, if that's a problem, and people who don't 
know how to type them would only have to learn the code to input, rather than 
the unicode stuff, or whatever).  Anyway, either way is fine with me, if you do 
it.

Anyway, thanks for letting me post, and do well, with whatever you do.