[math] [Fwd: Re: First proposal for COLT]

["Mark R. Diggory" <md...@latte.harvard.edu>]

Thought this might be of interest to the group.



-------- Original Message --------
From: - Sat Oct 25 13:03:08 2003
X-Mozilla-Status: 0001
X-Mozilla-Status2: 00000000
Message-ID: <3F...@latte.harvard.edu>
Date: Sat, 25 Oct 2003 13:02:54 -0400
From: Mark R. Diggory <md...@latte.harvard.edu>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; rv:1.5) Gecko/20030925
X-Accept-Language: en-us, en, zh, zh-cn, zh-hk, zh-sg, zh-tw
MIME-Version: 1.0
To: Discussion about JPackage project <jp...@zarb.org>, 
paul@honeylocust.com
Subject: Re: First proposal for COLT
References: <10...@urfida.trippa.org>
In-Reply-To: <10...@urfida.trippa.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Please excuse the crossposting, I think this would be interesting for
several different groups (including Paul Houle, the developer of RngPack).

This may be of some interest to you. There's been some effort at the
Jakarta Commons Math project to review much of whats happening with
COLT. We are doing our own development, some features of which do
overlap with COLTs capabilities. We are currently in the process of
"voting" to get out of the commons sandbox and into commons proper at
jakarta apache, we will be releasing and hopefully this will also
include rpm packaging at jpackage.

I've also spoken with Paul Houle and he has been open to relicensing
RngPack under an apache license so we can repackage it as part of the
jakarta math library as well.

I would like to consider the following idea, eventually we had ideas to
produce a "generic" api for math functionality (Random number
generators, statistical functions, linear algebra api's) that could be
implemented by any vendor, of course we're on a long way down this
particular road. However, efforts to rpm package COLT provide an
interesting opportunity to explore some avenues of these thoughts of
modularization we are having at the math project. With this in mind I
would enjoy being involved with some of this effort to rpm package COLT
and to also consider how Jakarta Commons Math and COLT may conflict or
interact with one another in the long run.

-Mark


Sebastiano Vigna wrote:

> OK, it took me several hours of digging, but here's the deal to install
> COLT correctly:
> 
> 1) A rngpack package, containing the random generators from RngPack.
> This had to be patched so that RandomElement is cloneable and has a
> clone method that simply delegates to parent, as code in COLT uses
> .clone(), but it does not work any longer in separate packages because
> clone is protected. This is GPL.
> 
> 2) A concurrent package, containing the concurrent stuff from Doug Lea.
> This is public domain.
> 
> 3) A violinstrings package, containing the ViolinString stuff. This is
> more-or-less LPPL (you can redistribute, but modification must change
> filenames).
> 
> 4) A corejava-format package containing the Format class from corejava.
> Whereas the remaining classes are non-free *this* specific class has
> been rereleased under the LGPL.
> 
> 5) Finally, a colt package with all the rest. It required several
> tweaks, but it works.
> 
> The colt package contains 1) the COLT stuff 2) the VNI math libraries
> and 3) the JAL libraries. The reason why things are set up in this way
> is that the sources for this libraries are very old and can apparently
> be found ONLY in the COLT distribution. 
> 
> Of course, an alternative is to package separately even the sources of
> those files, and distribute them somewhere.
> 
> There are however several things that are at best difficult to fix: COLT
> has an extensive introduction and tutorial that quotes the various
> packages of the distribution. That should be edited to make it coherent
> with the new distribution. But I do not think that it is nice to modify
> someone else's text.
> 
> Another problem is caused by the "abuse" of -docTitle made by the COLT
> author: it feeds javadoc with a -docTitle that is essentially a whole
> document body. The current version of javadoc (Sun 1.4.2_02) chokes
> and gives a bad output. I patched the overview file to get a sensible
> result.
> 
> There is still some polishing to be done, but I would like to know if
> the idea is OK.
> 
> Ciao,
> 
> 				seba
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> JPackage-discuss mailing list
> JPackage-discuss@zarb.org
> https://zarb.org/mailman/listinfo/jpackage-discuss

-- 
Mark Diggory
Software Developer
Harvard MIT Data Center
http://www.hmdc.harvard.edu



-- 
Mark Diggory
Software Developer
Harvard MIT Data Center
http://www.hmdc.harvard.edu


---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org