This is the mail archive of the guile@cygnus.com mailing list for the guile project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: guile: going the way of DEATH






  Many interesting comments have been made on this subject. I hope Jim
  Blandy and others are not discouraged by them.

  I agree that some parts are a mess. The documentation needs updated
  as soon as possible. Without good docs, you have to really want to
  use Guile to work with it. This will scare off the casual programmer
  that is looking for an extension language.

  Why am I using Guile? I believe that Scheme is by far too good of a
  language to go to waste. It just kills me to see so many concurrent
  Scheme developers working on separate projects - SCSH, MzScheme,
  RScheme, STk, SIOD, ELK, and Guile to name a few.

  It must be due to the nature of Lispy programmers, you don't see a
  half-dozen PERL or TCL variants being concurrently developed. Oh
  yes, I realize that is the point of RxRS, to unify the dialect. The
  problem is, RxRS is only suitable for teaching programming concepts
  and trivial "guess the animal" programs.

  As soon as a Scheme dialect is extended to do something *useful* it
  becomes incompatible with practicaly every other Scheme in
  existence.

  This is not a rant. I dearly want to see Scheme succeed as a
  scripting/ extension language. I also want to se it succeed as a
  stand alone programming language. I have written many mission
  critical PERL programs just because there was nothing more elegant
  to use (I simply refuse to use TCL).

  I feel that in order for a language to be really useful for
  scripting, extension, *and* development it needs:

      1. Minimal startup time. Yes, I know you can reduce Guile's
         startup time by various hacks. It *needs* to be fast right
         out the box.

      2. Minimal core. Everything other than basic functionality
         needs to be a module.

      3. Drop dead simple networking (a module). Guile more or less
         does this already, except that it needs to be a module and
         recv! needs supplemented in some way.

      4. Drop dead simple GUI (a module). Like STk. I know GUI isn't
         important to everyone. Add it to the above features and you
         have a language that can compete in the Real World with PERL
         and TCL.


  These are my perceptions of the matter, yours will probably be different.


  -Dave

--
David Tillman            |    Sparrow Information Systems
dtillman@sparrowsys.com  |  Contract C and PERL programming