Re: Scheme tutorial

["Ian Bicking" <bickiia@earlham.edu>]

>  > Opinions on how much I should emphasize scripting an application
>  > vs. writing a more generic piece of Scheme code?  Is there any
>  > application that might be considered a reference work for scripting?
> 
> unabashed self-plug: i volunteer thud, mainly because it needs bugs
> shaken out, but more importantly, its API needs to be more extensible
> (and finalized for 1.0).  i envision your tutorial could take as a
> case-study a simple 8-bit uproc to implement...

8-bit uproc?  One problem with Thud is that, well, I don't really 
understand the problem area of chip simulations, except maybe 
enough to realize that's what Thud is for :-)

>  > I.e., would there be a Guile-scriptable application that could be
>  > considered a prerequisite for working through the tutorial, or is
>  > that too much to ask from the reader?  Installing any particular
>  > application can be a pain, and I'd like to make the tutorial as low-
>  > commitment as possible.  OTOH, I'd like to have a tutorial where
>  > people can do some neat things early on, and scripting an application
>  > is a good way to do this.
> 
> what you're looking for is a high-level toolkit rather than an
> application.  (either that, or an application that is sufficiently
> modular so that its toolkit is exposed.)  the way i see it, an
> application proper would not really require further scripting.

Well, I was thinking more along the lines of scripting, which is 
something much more casual than writing an application with a 
toolkit.  It is possible that I could do a bit of both -- starting out with 
some scripting and moving to something a bit larger.  Hmm...

An application can certainly need scripting.  Probably the simplest 
being a macro, the most complex being all those complicated elisp 
packages.  Maybe elisp starts turning emacs into a full-featured 
toolkit+environment, but that starts happening when the 
programmer is well past the skill level this tutorial would be working 
with.

In many ways Gimp would seem ideal, except that it doesn't use 
Guile (sigh...), it's big, and to script it you have to understand a fair 
amount about image manipulation.  I'm not sure why I thought it 
was ideal, now.  So much for that idea...

Thud has the problem that you need to know something about 
simulation of a certain sort.  The example app should be something 
simple and universally understood... a Guile-scriptable 
calculator...?  Or is that too boring?

Maybe a modular tutorial that could be adapted for many 
applications... hehe... here I am, projecting notions of code-reuse 
onto writing.  A bit too optimistic.


--
Ian Bicking <bickiia@earlham.edu>