[MEI-L] [SPAM] Re: Introduction and engraving/typesetting considerations

[Urs Liska]

Am 05.07.2014 14:01, schrieb Klaus Rettinghaus:
> Hi Urs and all,
>
> I started to put a mei2ly in XSL together already, planning to upload it in
> a few weeks.

Where do you want to upload that?

> Its a start, but that's surely a one way solution.

Of course we should combine any efforts done by all interested parties. 
And maybe it's also good to combine different approaches.

One thought: Creating a general ly2mei converter for arbitrary existing 
files will face the same problems as the MusicXML export we're working 
on. The problem with that is that the LilyPond input language allows 
more than a one-to-one representation like MusicXML. You can reuse music 
when stored in variables. OK, that can be treated. You can also reuse 
music in processed forms (transposition, rhythmic scaling etc.). OK, 
somewhat more involved but can also be handled. But using Scheme you can 
do all sorts of stuff to the content up to algorithmically or randomly 
generating the content during the LilyPond run. This can only be 
converted by redoing everything that LilyPond does - or by capturing the 
information during the LilyPond compilation.

Thinking about an approach with a kind of dual-mode editor I would 
restrict its functionality to what is reasonably possible to do with 
both formats and not try to convert arbitrary files. I.e. this would 
probably only work flawlessly with documents that are created for that 
purpose from the start.

> To achieve
> a better integration with Lilypond a good idea would be probably to fork
> the musicxml2ly module from Lilypond.

If so one should definitely also consider this existing fork 
https://github.com/Philomelos/lilypond-musicxml2ly-dev
which has a number of improvements that haven't been backported to 
LilyPond's own musicxml2ly.

But maybe it would be even better to go the other way. As mentioned in 
my paper we have by now a powerful library to work on LilyPond files 
through a DOM, and I think it could be promising to go that way, i.e. 
rather interface from one DOM to the other than  parsing and generating 
LilyPond code manually. Of course this may also involve XSL 
transformations on certain stages.

Unfortunately I don't know my way
> around in python.

Well, I don't have experience with XSLT ...

Best
Urs

>
> Klaus
>
>
>
> _______________________________________________
> mei-l mailing list
> mei-l at lists.uni-paderborn.de
> https://lists.uni-paderborn.de/mailman/listinfo/mei-l
>