Re: PATCH - DOC: Added @knownissue to NR for fingering

[David Kastrup]

Graham Percival <address@hidden> writes:

> On Sat, Mar 12, 2011 at 03:00:41PM +0100, David Kastrup wrote:
>> David Kastrup <address@hidden> writes:
>> 
>> > Here is one example where -3 and -13 do totally different things:
>> >
>> >
>> > xxxx=-3
>> > yyyy=-13
>> > #(display xxxx)
>> > #(display yyyy)
>> 
>> Incidentally: does anybody have a reasonable idea how we want to get
>> around this?
>
> completely untested:
>   yyyy = -1.0 * 13
> ?  :)

-13 is no problem.  -3 is...

> The serious answer is to add # in front of all numbers; we started
> doing this in the docs back in 2007.

You mean, not allowing numbers at all in contexts where they could be
confused with fingerings?  Rather than just not allowing one-digit
negative numbers?

>> Currently it would appear impossible to assign one-digit negative
>> numbers to variables since they are instead interpreted as fingerings.
>> 
>> That seems like a rather appalling deficiency in the syntax.
>
> Oh, agreed.  I think we should remove the DIGIT stuff.

It does seem to cause some peculiarities.  The question is what other
peculiarities we are buying ourselves instead.

Here are some amusing thoughts:

xxx=-5  # assign -5 to xxx

c-\xxx   # equivalent to c-5

The idea here would be to make -\xxx where \xxx is a numeric identifier
equivalent to using -\xxx as a fingering.  That would have the quirky
consequence of

xxx=5

c-\xxx

generate a fingering with a value of -5.

It would likely resolve the current situation, meaning current inputs
that are intended to be valid would remain or become so.

there is a possibility that this is "too cute".

-- 
David Kastrup