[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: brainstorming a really smart system engraver
From: |
David Kastrup |
Subject: |
Re: brainstorming a really smart system engraver |
Date: |
Thu, 28 Aug 2014 10:28:19 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/24.4.50 (gnu/linux) |
Janek Warchoł <address@hidden> writes:
> 2014-08-28 1:40 GMT+02:00 Kieren MacMillan <address@hidden>:
>> Hi all,
>>
>>>> I think that issue 3518 (pushed recently) does just this:
>>>> https://code.google.com/p/lilypond/issues/detail?id=3518
>>>
>>> It doesn't do the automatic "AI nightmare" part.
>>
>> Yes, unfortunately...
>
> Yes, sorry - i should've trimmed quoted email better.
>
>>> However, it provides the low level machinery for pulling in the
>>> "maximally required" number
>>> of staves between automatic or manual line breaks, where the requirement
>>> is determined by working with keep-alive-interfaces and tags on the
>>> various staff variants.
>>
>> That could be helpful!
>>
>> I still need to wrap my head around how this framework/machinery
>> works (or doesn’t) with
>> true content-presentation separation; the example on the Google Code
>> page has multiple
>> "\context Staff” calls buried in the \violins note definition, which
>> to my mind mixes content
>> with presentation in an unfortunate way.
>
> I think you looked at an earlier work-in-progress snippet - in the
> attachment you can find the "final" version.
Which is still oversimplified: this is a regtest possibly pretending to
be too much. The main problem with the regtest is that it typesets
everything in two versions whereas a "proper" version should replace the
single-staff variant with skips while the double-staff variant is known
to win: it's not just saving unnecessary processing time, but there is
no point in letting the larger horizontal space requirements of the
single-staff variant determine line breaks in passages where the
single-staff variant is known to be irrelevant.
That's why I pointed to the different unisono/divisi issue's attachment
which does a more thorough job.
--
David Kastrup