commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gilles <gil...@harfang.homelinux.org>
Subject Re: [ALL] The Commons Math issue
Date Wed, 12 Apr 2017 23:16:24 GMT
Hi Oliver.

On Wed, 12 Apr 2017 21:41:34 +0200, Oliver Heger wrote:
> Am 12.04.2017 um 19:39 schrieb Gilles:
>> On Wed, 12 Apr 2017 18:25:03 +0200, Emmanuel Bourg wrote:
>>> On 04/12/2017 05:29 PM, Gilles wrote:
>>>
>>>> Do you actually prefer advertizing a non-Apache project rather 
>>>> than
>>>> having the PMC support its own developers in any which way it 
>>>> could?
>>>
>>> If nobody is able to maintain commons-math I have no objection
>>> recommending an alternative, especially one that is derived from
>>> commons-math, has the same license and an open development process.
>>
>> The issue here is that an "in-house" solution has been proposed,
>> based on time-consuming work on the part of developers still
>> contributing here.
>> The PMC members should logically (?) favour any proper endeavour
>> that attempts to keep _this_ community alive.
>>
>> For functionality that requires expertise not existing anymore 
>> around
>> here, it would be fine though, of course.
>> Thus I ask that we make a list of such functionality before 
>> dismissing
>> the local goodwill as if it didn't exist.
>>
>>> The minimal support you can expect from the PMC members is people 
>>> voting
>>> on the releases, and if there is no show stopper like binary
>>> incompatibilities, awful regressions or improperly licensed code, 
>>> the
>>> vote will be a non-issue.
>>>
>>>
>>>> How can you be so sure? The last releases did not elicit an awful 
>>>> lot
>>>> of votes; and that is for components that do not raise objections 
>>>> about
>>>> their mere existence.
>>>
>>> Give it a try?
>>
>> OK for small, focused, components?
>
> I am fine with Commons RNG and Commons Numbers.

Good to know!

> I would feel uneasy with a significant number of mathematical 
> components
> extracted from [math] that are added to Commons, even if they are 
> small
> and focused. It would seem strange if you opened the Commons Web site
> and about half of the components were math-related.

How fortunate, then, that we have so few contributors! :-}

Seriously: There are 28 "Commons" components having had an official
release since 2014; there isn't the slightest chance that what you
worry about can happen.

My suggested top-priority goal would be to have "Commons Numbers"
released, with more modules and/or more code borrowed from CM:
   http://markmail.org/thread/j5532mnsrgu4jzkv

These include utilities that easily qualify as "Commons"-type (as
evoked by Eric) or natural "Numbers" companions (e.g. "MathArrays").

The other concrete proposal was "Commons SigProc", that qualified
for inclusion. [Unfortunately, we lost contact with Bernd Porr...]

Then we can release a "legacy" CM 4.0, with up-to-date dependencies
to "Commons Numbers" and "Commons RNG", and several other improvements
and bug fixes that could benefit some users of CM.

Gilles

> If this is the goal,
> I would prefer to start again the top-level-project discussion.
>
> Oliver
>
>>
>> Gilles
>>
>>
>>>
>>> Emmanuel Bourg
>>>
>>
>>



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


Mime
View raw message