Re: LilyPond Architecture explanation

[Carl D. Sorensen]

On 3/30/09 7:42 AM, "Han-Wen Nienhuys" <address@hidden> wrote:

> OK - there is a weakness in the description that it skips all of the
> formatting work, which is the more complex and more flexible part of
> lily actually. This is because Erik's subject only dealt with the
> frontend part of lilypond, and did not modify the backend at all.

I understand.

If I were to write something of equivalent length about the backend, what do
you see as the key steps that should be explained?

I'm somewhat familiar with a step of creating grobs, which I consider the
first engraving step.  I'm not sure exactly how/when callbacks are used in
the grob creation process.

The grobs are then somehow placed in provisional staffs and pages, which is
somehow the domain of the spacing engine.  I have no understanding of this
phase at all.  I'd appreciate any pointers about this structure.

Once things are properly placed, they're written to an output file (.ps by
default, but .svg is possible as well). If the ouput is .ps, then ps2pdf is
called to create the pdf file.

There you have my rudimentary understanding of the backend.  Anything you
can clear up, or any place you can point me to get a better understanding,
would be greatly appreciated.

Thanks,

Carl