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 Fri, 14 Apr 2017 19:31:44 GMT
Emmanuel,

On Fri, 14 Apr 2017 19:37:08 +0200, Emmanuel Bourg wrote:
> On 04/14/2017 06:12 PM, Gilles wrote:
>
>> [I can be clearer: I had this issue with Emmanuel, about the
>> design and scope of "Commons RNG" (cf. ML archive), and it was
>> acknowledged that "do-ocracy" must prevail over "opinion".]
>
> Well, I had much more than mere "opinions" to offer to RNG,

You may have had, but you only gave opinions.

> but you
> rushed

False.

Development started in December 2015 in the Commons Math
repository, with plenty of noisy messages (cf. reference in
previous post of this thread, for examples).

Then, after the CM fork, I made an appeal on this ML to let
me create small components, first among them: "Commons RNG",
on which I had working (openly, and on a daily basis) for six
months).

Every aspect of the new code, especially those that departed
from the design implemented in "Commons Math" classes, has been
the subject of JIRA issues, and posts on the ML. Reasons were
thoroughly laid out (to become the Javadoc and userguide).

Zero objection.

Then I requested a git repository from INFRA, and moved the
code from package "o.a.c.math4.rng".
In August 2016, after most of the work was done (based on the
proposed design, refactoring of the CM unit tests, adding more,
running external validation tools), I opened an issue to collect
the remaining tasks in view of a release:
   https://issues.apache.org/jira/browse/RNG-6

Most of the code had been stable for months.
But I worked on yet another refactoring in order to have
modules (since a part of the PMC opposed the creation of
a separate "Commons Sampling" module).

In September 2016, I intended to cut a release, and only
then you started to question some of the foundation of the
design, on the ground that you wanted the component to
deal with functionality that had never been within the
intended scope.

Fortunately, I got support from one or two people who
also agreed that "do-ocracy" should prevail. You yourself
admitted that you arrived late into the game.

When the release was nearing, you asked to delay it for two
weeks in order to implement one or two generators.
Months passed, during which I ported two (other) generators.

Tired of waiting for a sign from you, the release happened
in mid-December.

Thus it took one year to release 1428 relevant lines[1] of
code!

> and imposed your design.

You acknowledged that you did not follow the issue (questions
and discussions in the ML, JIRA issues, commits) and you were
asking questions that had been thoroughly exposed in the
that development history.
Even so, I answered to all your posts, but you were never
satisfied because you were actually talking about another
component, one about which my opinion was that it would be
a can of worms that would delay (ad infinitum) the release
of what was already available.

> So be it, I gave up, I didn't have
> enough free time to keep up and argue with you,

That's the point, you argued but did not write any alternative
code.

> and there are other
> areas where I can be useful.

Sure.


All the best,
Gilles

[1] https://coveralls.io/github/apache/commons-rng

>
> 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