commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matt Cliff <cl...@mattcliff.com>
Subject Re: [Math] common-math and bloated 3rd party libraries
Date Wed, 05 Nov 2003 21:25:13 GMT
On Wed, 5 Nov 2003, Mark R. Diggory wrote:

> 
> Hmm, not to be critical, but 5.1 is almost at the end of its product
> lifecycle. Weblogic has had several releases since 5.1 that solve many
> of these issues do they not? I say this mostly to "identify" that there
> is a limitation as to how far back in terms of versioning we, as a "new
> tool", should consider supporting. I would have attempted to deal with
> an issue like this with BEA, not neccessarily with the java community at
> large because they create too many jars (I'm saying this lightly).
> 
> Not to nip a subject in the butt. But this has moved way off the issue
> associated with how dependent the different commons projects (and
> specifically related to math dependencies). I hope we can draw it back
> into this subject.
> 
> -Mark
> 
> 

  I agree with minimizing the number of dependent Jar's.  I also agree 
with statements made earlier in this thread related to supporting older 
JDKs (if JDK1.2 works for your app why should use of this component force 
you to change?)

  Regarding the logging;  It has been suggested that the logging 
dependency be isolated to the test classes only on not required for the 
runtime lib.  This makes sense to me, with the caveat that we must have a 
robust exception hierarchy so that logging can be plugged in at a higher 
level.

  Regarding the Discovery, I see only two instances where the 
DiscoveryClass is used (UnivariateRealSolverFactory, and 
DistributionFactory), both of which use only the newInstance(Class,String) 
method.  Off hand I dont see the advantage of using the DiscoveryClass 
here,  my impression is the find() methods are the 'meat' of the 
DiscoveryClass.


   commons-collections,  this seems only prudent to keep, as I can only
imagine that the linear algebraic routines will become more complicated,
and collections lends itself naturally to mathematical constructs.


  this leaves commons-lang, and commons-beanutils
None of these are used very extensively, however I can see how these are 
useful.
   commons-lang is used only in the MathException class,   as a trade off 
for removing logging I think that the NestableException becomes necessary.

   commons-beanutils is used by the *Transformer classes (seems good to 
me).





-- 
      Matt Cliff            
      Cliff Consulting
      303.757.4912
      720.280.6324 (c)


      The label said install Windows 98 or better so I installed Linux.


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


Mime
View raw message