Re: order of function arguments

[David Kastrup]

Janek Warchoł <address@hidden> writes:

> Hi,
>
> On 2013/04/21 08:23:34, dak wrote:
>> Janek Warchoł <mailto:address@hidden> writes:
>>
>> > 2013/4/16  <address@hidden>:
>> >> Uh, a call of offset can be followed by music quite naturally.  So this
>> >> is not just a "technical restriction".  It is one of logic.
>> >
>> > So, do i understand correctly that \offset could have a syntax like
>> >
>> > \offset grob-property-path value music
>> >
>> > And then, if grob-property-path contained only property,
>> > the function would be used as a tweak (with "music" being
>> > what's tweaked), and if the function was being used as an
>> > override (grob-property-path contains a grobname), "music"
>> > would be sort-of-unused?
>>
>> Possible, but "sort-of-unused" is not a reasonable interface.
>
> Hmm.  What do you suggest then?

Uh, what we are using already?

> It's not doing us good when user interfaces for different
> functions don't have a common specification wrt/ argument
> order.

Where do we have different argument order to existing functions?

> Do you see any reasonable strategies for solving this?

This was issue 2929
<URL:http://code.google.com/p/lilypond/issues/detail?id=2929> and I
answered your question in comment #7 in October.  The only alternative
is to provide _only_ overrides and force the user to use \single if he
needs a tweak.  But this requires the user to know the grob name for the
override, defeating the simplicity of \tweak.

> (as you know, my ultimate goal is merging as many modifying
> commands, e.g. merging \set and \override - but i'm not trying
> to suggest any solutions, because i have no expertise in this
> area, just some expectations).

I don't see a better solution here.  Obviously, or I would not have
converged to the behavior implemented in October.  Unless you can come
up with anything that has not already been considered and/or discussed,
this is not going to lead anywhere.

-- 
David Kastrup