I think autocorrection like that is a"feature" which I'd normally
describe with a three letter word that begins with B and
rhymes with rug. (In IBM we used to describe such bugs
as "working as designed". We later changed it to "Broken As
Designed").
I think that we need a rule - it isn't good enough to say that
ABC might mean all octaves or might not. The rule that
the first occurrence of a letter applies to all octaves and
following occurrences (if any) of the same letter apply only
to the given octave is the one I'd go for. The only down-sides
I can see are
A. The rule is complex - but you only need to face the full
complexity when dealing with tunes that need it. Otherwise
"keysigs apply to all octaves" is good enough.
B. K:D ^G =G
looks odd. Of course it means "sharpen Gs in all octaves
except the =G one". I think this is just a quirk. I will
probably go forever without using it. My guess is that to
anyone actually wanting that scale, it would feel like a
perfectly good way to write it.
Laurie
----- Original Message -----
From: John Chambers <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Thursday, February 01, 2001 6:18 PM
Subject: Re: [abcusers] developers/user
> Phil Taylor writes:
> | There is a small ambiguity which affects both global accidentals
> | and accidentals added to the key signature. In conventional notation,
> | sharps and flats in the key signature apply automatically to all
> | octaves, whereas true accidentals written before a note in the tune
> | apply only to that note (and any others of the same pitch for the
> | duration of that bar*). Notes of the same name, but in different
> | octaves are unaffected. Clearly, if the proposal is to deal with
> | exotic scales which differ between octaves, we need the extra
> | accidentals which are added to the keysig to operate like true
> | accidentals, leaving notes of the same name but in different octaves
> | unaffected, rather than like the normal components of the keysig.
>
> Well, I'd argue that conventional staff notation doesn't actually
> deal with this issue at all, and neither should ABC. For music
> formatters like abc2mtex and abc2ps, it's a non-issue, because they
> don't produce pitches. They only need to draw the key signature, and
> the interpretation is up to the reader. If different gangs of
> musicians want to use different rules, fine. This doesn't affect the
> notation at all.
>
> However, unlike printed music, ABC is also fed to players such as
> abc2midi. These programs do produce pitches, and they need to know
> how to interpret the key signature. The obvious approach is to give
> the user an option or menu or bunch of checkboxes and let the user
> decide. The default should probably be the usual interpretation of
> applying the keysig accidentals to all octaves. Or maybe this should
> be the default unless the key signature contains explicit accidentals
> on the same note in different octaves, in which case it gets more
> complicated.
>
> | Or perhaps the rule should be that an accidental in the keysig
> | operates on notes of all octaves unless cancelled by a subsequent
> | accidental, and only the second accidental is specific to one octave?
>
> Maybe ABC's rule should be "It's Someone Else's Problem". ;-)
>
> A friend just discovered an interesting bug/feature in abc2midi: He
> had a tune in K:G, and in one bar there was an F harmonized by "F".
> What came out the speaker was F natural, both in the melody and the
> chord, although he hadn't written =F for the note.
>
> He thought this was a bug, because he had been using the midi version
> to proof the ABC, and since it sounded right, he didn't spot the
> missing '='. I pointed it out to him, which led him to investigate.
>
> We verified that this only affects the root of the chord. It gave the
> "F" chord the correct interpretation of an F major chord, and then
> modified the melody F (which should have been ^F) to match the
> chord's root. But a test with "A7" for the chord and a c in the
> melody didn't modify the c to ^c.
>
> It's a somewhat related topic, in that the software is doing an
> "intelligent" job of noticing the transient accidental implied by the
> chord, and applying it to the melody even though not written in the
> ABC. Interesting ...
>
> To subscribe/unsubscribe, point your browser to:
http://www.tullochgorm.com/lists.html
>
To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html
