Jean-Francois Moine <[EMAIL PROTECTED]> mentions:
| John Chambers a skrivas:
|       [snip]
| > before or after a |. The abbreviation :: may be used for :|:, i.e., a
| > combined end-repeat and start-repeat in the middle of a staff.
|
| I see ::, but no :|: in the actual standard/draft. Do we have to
| implement it?

Good point.  Something that has been mentioned before that I think is
a good idea:  Redundant bar lines should generally be reduced without
comment.  This is already done in some cases by abc2ps.  Thus, if one
line ends with :| and the next starts with |:, and you've used the -c
option (meaning to ignore line  endings  and  generate  staff  breaks
automagically),  the  pair  is reduced to ::.  Similarly, ::  will be
rewritten as :|\n|: if that's where a staff break ends up.  There are
a lot of abc files around with endings like :|| and :|], which aren't
strictly legal by the current standard, but probably should  be  made
legal.  I saw a case of :|: earlier today in another mailing list. So
:|:  and :||:  should be accepted as meaning the same as  ::   (which
should probably be considered an abbreviation).

The best justification is the old advice to be liberal  in  what  you
accept  and conservative in what you produce.  The other form of this
argument is that it makes life easier for  programmers  in  the  long
run.   A  lot of abc has been and will be generated by software.  The
programmer's job is made a lot easier if you can do things like  just
produce  :|  at  the end of a repeat without the routine knowing what
comes next.  This will tend to  produce  :||:   in  the  naive  case.
Requiring the program to reduce this to :: is a needless complication
that wastes many programmer hours. The abc can't just be produced; it
must  be  accumulated  in memory and examined for possible rewritings
like this. It's far more efficient to just output the simple abc, and
then chew these things up during input and do the reduction there.

There's also the fact that abc isn't just a  computer  code.   It  is
written and read by humans. Many of them are musicians with very poor
understanding of the needs of software.  Current practice shows  that
they do generate all sorts of "reasonable" abc.  Life would be easier
for everyone if the software were also reasonable about such  things.
You  may  think that it makes extra work for the programmer, but that
bit of extra work saves everyone a lot of hand-editing to correct the
abc  from  non-expert  users (or users of software that produces such
non-standard abc).

| >    |:   ... | ... |[1,3 ... :|[2,4 ... :|
| >    |::: ... | ... |[1,3 ... :|[2,4 ... :|
...
| Right, over 45 years of my music life, I never saw such a notation!

Where've you been, boy? I've seen it a lot. I would say, though, that
it's  notation  mostly used and liked by dance musicians.  One of the
real problems with a lot of printed music is following  the  repeats.
I've  seen a lot of cases of a piece of music falling apart and dying
because  the  musicians  can't  find  the  start  of  a   repeat   or
misunderstand  how  many  times  to play a phrase.  If you're reading
music that has been put up on a stand, it helps a lot if the  repeats
are  VERY  clearly marked.  So experienced dance musicians will write
the |::: to give the reader advance warning that this is the start of
a  four-times  repeat.   The  bar line(s) will typically also be very
thick, and together with the three colons, it gives you a mark that's
easy  to  spot in a fraction of a second.  Most musicians not working
under such pressure don't know or care about such things, but some of
us  know  about  it  and  consider  it valuable.  And it's so easy to
implement ...

| OK for me, I will just ignore these multiple colons instead of adding
| dashed bars! BTW, could these last ones (simple colon) be the new
| standard?

Not sure what you're saying here.  In a case like |::  ...  ::|,  the
colons are not extra.  This means three times through, and there's no
other common notation for it. Ignoring the extra colons in such cases
is simply wrong.

To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html

Reply via email to