Re: Variable weirdness

[Simon Albrecht]

On 07.11.2016 17:56, David Kastrup wrote:
> David Sumbler <address@hidden> writes:
>
> > I had the following lines in the main file of my current Lilypond
> > project:
> >
> > \book {
> >    \bookOutputName "../firstCello"
> >      partName = "Cello 1"
> >      \include "frontcover.ily"
> >      \bookpart { %music...
> >
> > The file "frontcover.ily" contains a \bookpart block which prints a
> > front cover with title, composer etc. - these are defined elsewhere.
> >   But it needs one more variable, viz. 'partName'.
> >
> > I discovered that Lilypond will not accept a variable definition in
> > Lilypond format in the position I have put it: at the top level of a
> > \book block.  Nor will it accept it in a \bookpart block.  But at a
> > higher or a lower level, it will.
> >
> > This seems a bit weird (to say the least), in view of the fact that by
> > replacing the line
> >         partName = "Cello 1"
> > with the Scheme form
> >         #(define partName "Cello 1")
> > everything works as intended.
> >
> > Is there any useful reason why a variable cannot be defined in Lilypond
> > format in these contexts?
> Because they would not be local to these contexts?

In other words, to make this work, books and bookparts would need their 
own namespaces, am I right? What would be the drawbacks of that? I 
daresay it would be pretty intuitive to use, and I’ve also found myself 
wanting that feature in the past, mainly to simplify \include structures.

Best, Simon