Re: tuplet statsŠ

[David Kastrup]

Carl Sorensen <address@hidden> writes:

> On 9/25/16, 2:46 PM, "Simon Albrecht" <address@hidden> wrote:
>
>>On 25.09.2016 22:06, Trevor Daniels wrote:
>>> When needed, isn't this adequate?
>>>
>>> \version "2.19.48"
>>> triplet = \tuplet 3/2 \etc
>>> \triplet { a'4 a' a' }
>>
>>That¹s what I meant by Œimplementation is a no-brainer¹ :-)
>>Best, Simon
>
> I¹m generally opposed to syntactic sugar.  It tends to hide behavior that
> is important to know to move beyond the basics.
>
> I guess that this kind of discussion is what GLISS was supposed to be
> about, way back when it was planned.
>
> I would not be in favor of adding \triplet to the LilyPond distribution.
> But others may feel differently.

LilyPond input is not MusicXML.  It is supposed to be nice to humans.
My main beef is against arbitrary extension of LilyPond's syntax for
minor improvements, but this does not extend LilyPond's syntax.  It's
just a trivial abbreviation.

Personally, I'd not even think of it in terms of GLISS but then as
someone with a focus on implementing stuff I might draw the lines
differently than others: when introducing optional arguments and
fraction arguments and converting a whole bunch of stuff like \time,
\times (in pre-\tuplet times), \clef and so on into music functions,
that felt like removing them from the scope of "input syntax", the "IS"
in "GLISS": their respective input forms stopped having to make global
sense but could be chosen individually.

\triplet saves very little input, it does not actually cover all
triplets (like the rare 3/5 triplets) and it does not generalize well to
things like quintuplets which can really be 5/4 or 5/3 depending on
metre.

So I'm a bit meh on it even though I don't consider it GLISSful.  Its
cost is small but I don't see it as overly helpful either.  But then I
am writing on a US keyboard layout where 3/2 can be typed with 3 key
strokes and no modifiers (Shift, AltGr or whatever).

-- 
David Kastrup