Mogens Lemvig Hansen <[EMAIL PROTECTED]> wrote:
>Thank you for the reply - boy, I hit a nerve there :-)
>
>Thanks for clearing up my misunderstanding about who decides on line
>breaks in MusixTeX. I am also more comfortable that TeX decides the
>line breaks.
>
:-)) Sorry about the appearance of "raw nerves"... more just nervousness...
The last time I touched Musixflx I swore I would never go back again... I
might really screw somthing up. :-)
Anyway, what I wrote in haste on the way out of the house this morning is
not quite the whole story...
My previous response was somewhat oversimplified (and in part, not
correct!), ...
Musixflx really does execise control over where line breaks occur. For
example, by setting \linegoal one can illustrate the extent that Musixflx
really drives the line breaking decisions. However, musixflx makes all line
breaking decisions under the ASSUMPTION that it knows exactly what MusixTex
will do with each of a number of contingencies it finds during during Pass
2. It is supposed to be just a slave to MusixTex.
For example, musixflx assumes that if a sign change occurs in the middle of
a line, there will be no cautionary sign change just before that bar.
However, if it occurs at the end of a line, then it knows that MusixTex will
post a cautionary sign change at the end of the line, just before the next
line.
It was because musixflx did not know how MusixTex was handling sign changes
posted during xbars (when then happened to fall at the end of a line), that
it was once miscalculating the spacing for the two lines involved.
So, given a number of differenct contingencies (special cases) musixfix
currently knows about, the real problem in designing something like
\mudiscretionary is for the designers and maintainers to decide just how
much MORE musixflx will need to be taught in order to do the right
computation when the discretionary material falls in different places.
Taken to the extreme, musixflx would start to become more like a TeX
emulator. Because this was not the intention when musixflx was first
designed, any movement in that direction, without extensively redesigning
musixflx will simply add to the the "special cases" that already obscure the
situation too much.
This is why I strongly advocate avoiding more situations where musixflx
would need much more knowledge of the consequences associated with music
elements occuring in different contexts... (e.g.) end of line rather than
beginning or middle.
However, this is just my opinion, having tramped around in Musixflx more
than I ever wanted to! :-)
>
>My current problem is changing metres. I use
>\generalmeter{...}\changecontext
>which does what I want if the change happens in the middle of a line.
>However, if the metre change happens at a line break, I would like the
>new metre to be posted twice: at the end of the old line and at the
>beginning of the new line. That is:
>
>\beginpiece
> 4/4 music in 4/4 ...
> music in 4/4 ... |3/4
> 3/4 music in 3/4 ...
>\endpiece
>
>I can achieve the look I want using hard line breaks (\alaligne), but I
>would rather that TeX decides the line breaks instead of I.
>
>I have posted this problem before and then got the answer that it is
>impossible. I then figured that a \mudiscretionary command would be a
>generic solution to this type of problem. Is there another way? Or a
>way of implementing a \mudiscretionary command without causing trouble
>for musixflx?
>
>Regards,
>Mogens
>
Maybe a MusixTex-Pert could look at into MusixTex to see if the same logic
that applies to sign change posting decisions could also be (or is supposed
to be?) applied to Meter change postings. If that decision could be
instrumented within MusixTex, then Musixflx might be up for (say) one more
little special case. :-)
Joel Hunsberger
[EMAIL PROTECTED]
