RE: Fret Diagram markup

[Carl D. Sorensen]

> -----Original Message-----
> From: Han-Wen Nienhuys [mailto:address@hidden
> Sent: Tuesday, May 25, 2004 2:20 PM
>
<SNIP> 
> This was also an opportunity for me to take a closer look at
> your code. I still have some remarks:
> 
>  * We want to separate formatting machinery from syntax.
> Your fret-diagram command accepts a string containing items 
> separated by semicolons.  It would be better style to put the 
> items in a Scheme list, eg.
> 
>   \fret-diagram-verbose #0.75 #'(6-x 5-x 4-o 1-2)
> 
> or even better,
> 
>   \fret-diagram-verbose #0.75
>     #'((cross 6) (cross 5) (open 4) (barre 1 2))
> 

I'm delighted to make this change.  Thanks for the input on how you'd
like it to work.  Do you think it would be best to make a
\fret-diagram-verbose in addition to the current \fret-diagram, or
should I just make the syntax change to \fret-diagram?


> A separate frontend could translate the current string syntax
> to a Scheme list. 
> 
>  * The diagram items include settings for linethickness, height,
>    width, etc.
> 
>    It would be better style to put these settings into properties, so
>    you could do
> 
>          \translator { \StaffContext
>            \override FretDiagram #'thickness = #0.2
>          }
> 
>    and change all fret diagrams in one fell swoop.

    I assume that size should also be part of the properties, rather
than an argument to the markup.

> 
>    The properties delivered to the markup command through the props
>    argument (which is a list of alists)
>

So do I assume that props is provided by the context, and that I can
just get the value out of props with appropriate (assoc * ) calls?  Can
you give me some suggestions as to where I'd go to find good examples of
working with props?
 
>  * for multi-choice alternatives, such as
> 
>          f:number -- set fingering label type (0 = none, 1 = in circle
>                   on string, 2 = below string) Default 0
> 
>    we use symbols instead of numbers. They are more descriptive.

Of course this is the best way.  Should I just define symbols (define
BELOW 2) as is done for axes in calls to ly:stencil-translate-axis or is
it better to just use something like (if eq? finger-code
fc-label-below)?

> 
>  * For the same reason, I frown on single character command
>    names. Write
> 
> 
>      thickness, barre, radius
> 
>    instead of
> 
>      t, c, d
> 

No problem.  I can easily see the advantage.  And that's a reason why a
list as an argument is much better than a string, because I don't have
to parse, right?

>  * for alists, we generally use symbols, not strings as keys.

So intead of (acons "thickness"  thickness), I should use (acons
'thickness thickness)?

> 
>  * Unlike LISP, Scheme does not have nil pre-defined. Simply use '()

Oops -- I'll fix it.

> 
>  * It would be nice if the width of a diagram were computed, from the
>    length of the stringTunings property.

Which I assume is available from props via something like (assoc (car
props) stringTunings)?

> 
>  * linethicknesses are generally expressed relative to the
>    linethickness \paper variable. The default usually is 1.0
> 
> 
> I should be grateful if you could address these issues. 

I'll be happy to work on them as I can.  I'm traveling over the next
three weeks, which could be either good or bad in terms of having time
to work on them.

I appreciate your review of my code, and I look forward to being able to
increase my contribution.  Thanks for being willing to answer questions.

Carl