Re: Unconventional score and unwanted stray staff lines

[David Bellows]

Hey Kieren,

> If the algorithm uses a random number generator to determine which output to 
> give, “one never knows what’s going to happen”, but it would still be easy 
> for Lilypond to make the result "look good".

Yep, that's it exactly. I do spend a lot of time generating a lot of
different runs with different parameters and such looking for weird
gotchas and then coming up with general solutions when those things
happen. So far the results are really good which is a testament to
Lilypond's quality. The final result will be online and there will be
no chance for the user to fix the Lilypond file so I need to make sure
things look as good as possible.

Dave

On Sun, May 19, 2019 at 2:50 PM Kieren MacMillan
<address@hidden> wrote:
>
> Hi Mark,
>
> > You are asking a computer (Lilypond) to look good when "one never knows 
> > what's going to happen."
> > IMHO, the two are mutually exclusive.
>
> Doesn’t that depends ultimately upon the precise algorithms (and thus output) 
> involved? I mean, I could code an algorithm that outputs either
>
>   { c'1 }
>
> or
>
>   { c''1 }
>
> with no other possibility. If the algorithm uses a random number generator to 
> determine which output to give, “one never knows what’s going to happen”, but 
> it would still be easy for Lilypond to make the result "look good".
>
> Obviously this is a vast simplification of the actual situation under 
> discussion… My point is simply that the two constraints, while potentially 
> difficult to satisfy simultaneously, aren’t [literally] mutually exclusive.
>
> Cheers,
> Kieren.
> ________________________________
>
> Kieren MacMillan, composer
> ‣ website: www.kierenmacmillan.info
> ‣ email: address@hidden
>