Tobias Kunze wrote:

>  I think you are calling for a completely new synthesis
>  language.  While csound may be "ok" as a Music V-style
>  instrument description language on a basic level (ie, < 20
>  different unit generators), its expressiveness as a
>  "programming" language is damn close to zero.  
...
>  The "language" itself is a catastrophe:
>  you can't mix cases or pick meaningful variable names, are stuck
>  with table and instrument numbers instead of names and thus
>  can't embed code, can't print diagnostics or debug, and so
>  forth.



  I was wondering whether it would be possible to alleviate
  some of the shortcomings of the Csound language by 
  building instruments in a more friendly and ``higher level''
  programming language, then using a ``translator'' of some sort 
  to change the given orchestra into csound format.  Thus, the
  .orc file would be a ``low-level'' intermediate stage, much like 
  the object files created by C compilers.

  (Indeed, as a first embryonic step, it is not too hard to imagine a 
   ``sed'' script to change long upper-and-lower-case variable names 
   into Csound compatible names using a one-to-one association).

  This approach would eliminate the limitations with the variable name
  syntax, and possibly allow more modern flow-control conditionals.

  My question is:  Is the Csound language sufficiently
 ``complete'' to support most modern logical constructions (even
  if implemented in some sort of ugly way), or would such a 
 ``higher level'' language be simply a cosmetic improvement?



  - Fabio Bertolotti