RE: [Axiom-developer] Request for testing (Win32 users)

[Bill Page]

On February 11, 2007 1:22 PM Vanuxem Grégory wrote:
> 
> I have uploaded a version of Axiom built on top of SBCL for 
> Windows (32 bits). So if you are a Windows user, you already
> know Axiom and you can spent some time for testing and reporting
> bugs I would really be happy.

Great! I am very glad you are working on this.

> 
> Things that you have to know:
> 
> * You must know Axiom,
> * You have an official version of Axiom (built on top of GCL) 
>   for Windows,

I do, I do. :-)

> * You must know that this version comes with absolutely no 
>   warranty of any kind,
> * "How to use it" is not supported,

Isn't that already obvious and clear? I think this applies to
all open source projects including the open source version(s)
of Axiom.

> * Bug (that is, bug in this version not present in the 
> official one) MUST be reported to vanuxemg at yahoo dot fr
> and not to the official Axiom bug tracking systems,

Do you prefer this for your own personal reasons? I do not see
why your work should not be included along with the other current
and experimental version of Axiom. It is possible to make special
provision (i.e. new categories) in the Axiom wiki IssueTracker
system to refer to your new version so that we can easily
distinguish to which version the issue applies.

> * This is not the same version than the official one (the 
>   code differs),

I do not know what is really "official" any more. Juergen Weiss'
experimental version of Axiom on CMUCL has been discussed before
on this list and his original source code has been distributed.
Some of his changes have been incorporated into the "official"
sources in available via CVS and at SourceForge.

> * This is not a fork (code in this version is destined to be 
> included in the official version),
> * The source code is NOT available (sorry, not yet),

I think it would be perfect if you would make this source available
as a branch in SVN at SourceForge.

> * Do not discuss issues related to this version on the official
>   Axiom mailing lists unless this is explicitly permitted,

As Tim Daly has already pointed out: no permission is required
to discuss any Axiom related issue on the "official" Axiom mailing
lists. In fact, discussing Axiom any other way seems rather
pointless to me.

> * This version contains a LOT of bugs,

I think that this is not a surprise and does not differ from other
current experimental versions of Axiom.

> * This version is based on an "experimental" version of SBCL so
>   you'll contribute to SBCL too if you find SBCL bugs, but do
>   know that all of this is experimental,

Great. I am very glad that someone has finally begun to work
publicly with Axiom on another version of Lisp.

> * Do not trust its results.
>

That applies to ALL computer algebra systems - even the expensive
commercial ones.
 
> 
> It can be downloaded from:
> http://perso.orange.fr/parabola/axiom-sbcl-win.tar.bz2
> 
> To use it, untar/unbzip2 it in a directory without space in its
> name (full path without space). Export the environment variable
> AXIOM, it must point to the directory that contains the algebra,
> bin etc... directories.
> 
> So for example:
> 
> set AXIOM=c:\somewhere\axiom-sbcl-win
> 
> and you can start it with
> 
> c:\somewhere\axiom-sbcl-win\bin\interpsys
>

Thank very much for making this available.
 
> 
> One of the most important thing to know is that this work 
> is based on the work of Juergen Weiss (among others of course),
> he "ported" parts of Axiom to the CMU Common Lisp system.
> 
> Many, many thanks to you, Juergen Weiss.
> 

Yes, as I have said before: Juergen's work on the CMU version
was (and still is) critically important to the current open
source version of Axiom. I think it would be great if someone
would take on the job of making the source for the CMU version
available in the SourceForge SVN repository. I think there may
be some people who would be willing to test it further and to
contribute changes and improvements.

I don't think we should be afraid to work on several different
versions of Axiom in parallel, especially when one of the stated
goals is to of the Axiom project is to support Axiom on a number
of different Common Lisp environments.

Thanks again, Greg, for making your work available!

Regards,
Bill Page.