This is the mail archive of the guile@sources.redhat.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]

Guile IMHO and a proposal.


Before I begin this post I first wish to reference the replies to and 
resulting discussion from my last email that I sent to you and also posted 
to "guile@sourceware.cygnus.com" entitled "PHP fork project- Guile vs Python 
vs ?". Please don't read what I'm saying as FUD or flamebait, this is what I 
believe to be an objective observation of the status of Guile.

Essentially the resulting discussion and my own evaluation of the current 
development of Guile brought to light the following issues/limitations:

*Guile is an embeddable version of a subset some implementation of Scheme. 
-As there is little documentation this makes for tough work anyone wanting 
to use Guile even if they know another implementation of Scheme.

*Guile's website needs work. - This is an easy fix, from the quantity and 
quality of discussion in the Guile newsgroup all that is needed is a method 
of bringing the information from the newsgroup to the webpage. IMHO PHP had 
an excellent idea when they allowed users to post comments to their 
Documentation. That way any issues are posted to the topic rather than just 
someone else or a generic newsgroup. If anyone agrees with my evaluation the 
I'd be happy to work on this. The project I was working on before PHP 
decided to move away from the GPL is very much related to this. The only 
problem is that the program itself is a PHP script and would have to reside 
on a server with support for PHP3. If your current webhost doesn't support 
PHP3 I'd even go as far as to commit to paying for a account on CIHost to 
host Guile.org, setting the system up and maintaining it.

*Guile has little documentation. -Scheme has plenty of documentation but as 
this documentation is Scheme implementation specific it does little good to 
someone wishing to use Guile. See the above for a partial solution to this.

*Guile has no stable Module system. -Apparently this is the current focus of 
Guile development but is essential to extending Guile.

*Guile needs leadership - I know saying this is likely to piss people off 
but history has shown that for any project to succeed someone has to take 
the helm. This person needs to stand gatekeeper, provide direction and 
evaluate progress. Don't accept a change without the documentation of that 
change. Clearly set the focus for development and a roadmap to future goals. 
Honestly evaluate the current status of a project to the developemnt group 
as well as to the public. Often this will result in personality conflicts, 
especially in an environment where the main motivation isn't fame or 
personal wealth.

Guile must definetly be worth the effort though. I'm frankly shocked at the 
number of people working with/on Guile in light of the problems I perceive. 
This is evolution in action. Obviously people wouldn't be working on this 
unless they could see real value in doing such.

RMS's opinion is that Guile suffers from an inertia problem, but I'd more 
likely point the finger to a clear leadership problem. This isn't a slam 
again anyone in particular, just a general observation.

*Python->Guile is more feasible than Guile->Python. -I take this as somewhat 
of a compliment to Guile. If true, this speaks volumes about Guile's 
extensibility. Though this assumption may be based on comparing Python to a 
Scheme implementation.

Relating this to attempting a Guile PHP implementation, there seems little 
purpose in attempting such an effort. Once the module system is worked out 
and refined then considering implementing somthing similar to PHP as a Guile 
module or using Guile as a PHP interpreter may be feasible. Wether its 
practical is another matter entirely.

As I mentioned above, this isn't intended to be flamebait. I've outlined 
above where I would be willing to commit my personal efforts to the project 
if anyone else sees merit in my proposal. Aside from that I'll probably 
concentrate my efforts on maintaining a GPL'd version of PHP or more likely 
going back to working on my specification for yet another programming 
language (go ahead and wince, yet another idiot making a language in his own 
image). I don't make any claims about being the biggest brain or greatest 
hack but please consider my input.








>From: Richard Stallman <rms@gnu.org>
>Reply-To: rms@gnu.org
>To: tlowing@hotmail.com
>Subject: Guile and PHP
>Date: Sun, 9 Jul 2000 16:50:06 -0600 (MDT)
>
>     The only advantage I see to Guile from what I have seen is that 
>GNUCash is
>     (apparently) using it.
>
>A number of parts of GNOME use Guile.  Sooner or later, GDB will use
>Guile.
>
>			   Until mainstream apps start using it development will
>     be slow.
>
>Making PHP use it seems like a good way to make this happen;
>that's part of why I would like you to do it that way.
>
>     I could go on, but on the surface it appears that Python has many of 
>the
>     features that Guile wants to have.
>
>Features are secondary--Scheme and Lisp are fundamentally cleaner and
>more powerful than other languages, and no amount of added features in
>the other languages can compensate for that advantage.
>
>The fact that Python has more features than Guile today only means
>Python has a head start.  It would be sad to let Python beat Scheme
>for a trivial reason like that.
>
>Guile's biggest enemy is inertia.  I have made up my mnid that inertia
>is not going to win!

________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com


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