Re: commons-math and ActiveMath

[Paul Libbrecht <pa...@activemath.org>]

Le 10 sept. 04, à 05:06, Phil Steitz a écrit :
> I just poked about a bit on the ActiveMath site.  I could not get into 
> the CVS, btw.  It does not allow "anonymous" and "guest" wants a 
> password. Very interesting stuff.

Fixed, fixed... I find it partially a shame that maven 1.0 was released 
without even support for such overrides...

Just to explain this bit, JomeConsole is,  mostly, a wrapping:
- of Jome
	(http://www.activemath.org/projects/Jome/ actually a
       finished project of ESSI/INRIA Nice)
   providing the linear-parser to OpenMath (string)
- and a little evaluator using jdom and walking the jdom tree (of 
OpenMath elements) to evaluate the function

Jome is LGPL with main purpose to provide a "formula editor" for 
OpenMath. It's quite old, relying on an even older library of INRIA. 
Among the errors, I think, Jome is using as core model a linear-syntax 
model (which makes, e.g. a "+" a first class citizen).
Jome owners might, I believe, consider a donation at Apache... but it's 
main purpose is user-interface!

JomeConsole (including the evaluator) is MPL, could be donated with 
pleasure but is really small!

> I have a couple of questions for you.  First, would you be willing to 
> contribute / relicense, whatever, the expression parser that you 
> mentioned in your mail to commons-math?  I could not find the relevant 
> ActiveMath code in the downloads that I grabbed so I don't know 
> exactly what we are talking about here.  Could you tell me where the 
> relevant classes are (and if you would be willing to contribute them 
> to [math])?  If so, we can talk about it on the list. Actually best to 
> respond to this question on the list.

We had been exchanging with Mark a few months ago about where to find a 
package for formulae and there was nothing finished. One of the trends 
was Math-Beans, I think there's a little part in the source-tree 
currently.
I remember, at the time, already being annoyed by the dichotomy between 
complete generality (e.g. that matrices could be able to have entries 
in any ring) and... feasability. Currently, I think, what's in there 
considers more or less real numbers only, maybe complex numbers, but no 
fancy ring like Z_p or a ring of polynomials.

Mark was clearly stating that there's an amount of packages around that 
do formulae... (I remember now of Dautelle's JADE, 
http://www.dautelle.com/jade/) and that it was better to expect a 
donation from somewhere else...

Now this somewhere else hasn't come yet... maybe we should hunt and 
link around before ?

paul


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