commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gilles <gil...@harfang.homelinux.org>
Subject [Math] Not releasing 3.X (Was: Linear Programming in Math [...])
Date Tue, 02 Jan 2018 10:43:18 GMT
Hello.

On Mon, 1 Jan 2018 19:27:00 -0500, Rob Tompkins wrote:
>> On Jan 1, 2018, at 7:08 PM, Bill Igoe <billigoe@gmail.com> wrote:
>>
>> Hi Gang,
>>
>> I noted a recent exchange on OLS regarding Math Commons.  Thus far I 
>> find
>> the Math Commons working flawlessly.  I am also working on a project 
>> for
>> pure real time financial optimization using the Linear Programming
>> algorithm of Math commons.  I designed my code to flip between the 
>> QSOPT
>> package and the Math common LP algo to check the robustness and 
>> consistency
>> of  results.   I get exactly same answers in both and that LP is 
>> using over
>> 2000 variables and 3000 constraints!
>
> Many thanks for the encouragement.
>
> All - any thoughts on trying to do a release on [math] is the 3.X
> branch stable?

This branch is unsupported; making a new release based on it will
send the wrong signal and is likely to generate bug reports already
filed on "master" (with "Fix version" set to 4.0), and sometimes
fixed there (or in "RNG" or in "Numbers").

You'd basically scratch almost 3 years of continuous work:
---CUT---
commit e4e1ac23c734f65686be4bc0e503f82f941afd4d
Author: Thomas Neidhart <thomas.neidhart@gmail.com>
Date:   Mon Feb 16 23:37:23 2015 +0100

     Update for next development iteration: commons-math4
---CUT---

> Are there any small bug fixes we can do to make small
> incremental changes and release those?

Any time spent on back-porting fixes from "master" will be better
used for advancing towards the release of "Commons Numbers", and
other components with supported (reviewed and fixed) codes,
according to the "plan" (cf. ML archive for details).
Help welcome.

>
> -Rob
>
>>
>> Keep up the good work and I am looking forward to the 'split' 
>> between
>> Commons Math and Statistics.

As outlined in another thread, high-level functionalities with
positive feedback, like the one referred to here, can be ported
to a new component, with low-level supporting codes (but possibly
buggy) being hidden in "internal" packages until they are ported
(or released) themselves.
Comments, and help with experimenting, on this approach welcome.
E.g. the contents of "o.a.c.m.stat.regression" would become a
module of "Commons Stat".

>>
>> Cheers to you all and have a great 2018

Thanks and best regards,
Gilles

>>
>> Bill Igoe


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


Mime
View raw message