Re: [SPAM] Re: Transcribing individual score pages from scans

[David Cuenca]

On Mon, May 12, 2014 at 5:46 PM, Urs Liska <address@hidden> wrote:

What I haven't fully understood yet is how your users are entering content. Do they have an input field on a web page that they submit?

Yes, there is a text input field. For instance this page:
https://en.wikisource.org/wiki/Page:Childs_own_music_book.djvu/52

When you click "edit" becomes:
https://en.wikisource.org/w/index.php?title=Page:Childs_own_music_book.djvu/52&action="">

And there you can write the lilypond code wrapped with <score> and </score>

Then I'd repeat my suggestion. Let them enter parts of the content one by one, presumably providing them with user interface elements to define exactly what they are meaning (e.g. "the violin part for page 3").
If you can't use variables in LilyPond's safe mode then you might want to preprocess this input with a server side script. That is: store all fragments in individual text files with LilyPond variables, and before passing anything to LilyPond let some script generate a "flat" LilyPond file without variables.

I thought that maybe we could use comments to signal the different parts to a script to generat the "flat" version, but then I realized that there are many special cases, like "add lyrics" and I am unsure about how well it would play with Lilypond structure... It seems like forcing the sw to do something it is not prepared to..

About pre-parsing the variables before calling lilypond, I will have to ask the developer who wrote the extension to see if it is possible with the current pipeline.

Cheers,

Micru