On Mon, Oct 26, 2009 at 03:38:30PM -0700, Jeff Rogers wrote:

> The C-python implementation has a misfeature known as the global 
> interpreter lock (GIL) that presents scalability problems in a threaded 

Right.

> AOLserver is bound to tcl at a pretty deep level, and it does so very 
> well.  While there is some support for other languages (php, perl, 
> python, ML maybe?) it will never be as good as the tcl binding.

I'm not so sure your second statement there is true.  What I think IS
true, is that yes AOLserver is bound to Tcl at a pretty deep level, so
you certainly can't take Tcl out of AOLserver - nor should you want
to.  Tcl is pretty much in there to stay.

However, that doesn't necessarily mean that you can't make some other
language just as well supported for AOLserver application level
development as Tcl is, if you wanted to.  As far as I know there's
nothing in AOLserver's design that really prevents that, it's just a
Simple Matter of Programming.

Good choices would be any language implementation that's reentrant and
intended for easy embedding in C code.  Obvious candidates are
JavaScript, Lua, and probably a couple of the dozens of Scheme
implementations out there.

With various levels of completeness, people have tried:

- Standard ML, Objective CAML, PHP.  (I think all used in Production
  for limited or specialized purposes, e.g. to run canned PHP webmail
  apps.)
- Python.  (Sounded like a fair bit of development, but no live use AFAIK.)
- Scheme, Ruby, Perl, etc.  (Early development only AFAIK, never
  finished and haven't seen any real use).

See also:  http://panoptic.com/wiki/aolserver/Languages

In fact, although I am NOT particularly familiar with AOLserver's
internals (so I certainly could be mistaken), I'd bet that adding good
AOLserver support for some other scripting language would be easier
than in most other C-coded application servers which DON'T already
have tight and reasonably clean scripting language integration.

The Tcl support should have already laid most of the groundwork; you
shouldn't need to invent APIs and interfaces, just provide your new
Language X equivalents for the Tcl stuff that's already there.  The
only tricky parts might be language implementation specific magic like
the Tcl interpretor initialization.

I think a deeper question is, are you sure you want to do that anyway?
People who see AOLserver's usefullness usually just decide to use the
languages it comes with built in, Tcl and C.  But you could instead
try something completely different from the scripting-language-and-C
AOLserver-esque paradigm, like Erlang.  Or if you are a real hard-core
lover of language X, perhaps you want to build your own application
server yourself; I suspect that's what most language partisans or
enthusiasts do.

-- 
Andrew Piskorski <a...@piskorski.com>
http://www.piskorski.com/


--
AOLserver - http://www.aolserver.com/

To Remove yourself from this list, simply send an email to 
<lists...@listserv.aol.com> with the
body of "SIGNOFF AOLSERVER" in the email message. You can leave the Subject: 
field of your email blank.

Reply via email to