Re: critical issues

[David Kastrup]

Janek Warchoł <address@hidden> writes:

> According to our motto the aim of LilyPond is "music engraving to
> everyone" - i'd say it's a very good goal.  It would mean that a
> person with average computer skills (like navigating a web browser and
> using word processor) should be able to create very good engravings of
> not-so-complicated music (e.g. Mozart's "Ave Verum").  I think we're
> quite far from this goal; conductor of our choir didn't manage to
> switch to LilyPond.

So he did not manage to spell music in LilyPond in a few days.  I am
sure he tried to spell words in English for longer than that before he
gave up.

"Navigating a web browser" are not "average computer skills".  Those are
not computer skills at all.  You need more computer skills to program a
video recorder.

LilyPond is a batch processing system.  It is not an interactive
program.  You need to spell out the tasks.  People for which working
with a command line is an unsurmountable challenge won't work with
LilyPond.

They might get along with some GUI thingy that drives LilyPond as its
backend.

The one thing where LilyPond needs to refocus is getting away more from
the "fiddle around" stuff where you poke a program with a stick for
getting results rather than specifying your task.

But specifying your task in a computer-comprehensible way is not
avoidable.

In the areas of TeX/LaTeX and Emacs programming, I have hit a few
surprise candidates without programming background who put together a
significant body of macros and functions for their own work.  Pretty
much always it turned out that they were specialists in Oriental or
antique languages.

LilyPond is similar.  We can hope to get into a direction where it
requires less fiddling and programming skills, but writing things
directly in a computer language will always require logic, language and
thinking skills.

FORTRAN is a computer language in which good mathematicians can write
efficient numerical algorithms without tremendous programming skills.
As a result, there is a large corpus of numeric programming libraries in
FORTRAN.  It's not all that nice of a programming language, but it does
not add much of a distraction for a mathematician writing down numeric
code and/or using that of others.

If LilyPond manages a state where it does not get in the way of
composers writing down music, that is about the best one can hope to
achieve.  The tools and workflow can be made smoother, but that's it.

Don't _ever_ try to sell LilyPond to somebody who is not warm enough
with a computer to have a preferred editor.  In that case, you should
rather sell something like Frescobaldi.  Something which the user can
actually touch and see.

-- 
David Kastrup