On Sun, 20 Aug 2017 23:16:17 +0530, Amey Jadiye wrote:
I'm +1 to this change, I would be more than happy to lend my hands to
help
on this front. pardon me for being quite because some heavy work on
my day
job.
I don't want to fork this thread to different discussion but I have
one
simple doubt that rather creating whole new component why don't we
just
create maven modules within CM? I understand that release becomes
easy
with different component and some more other advantages but same can
be
done within single project. this is obvious question and which you
guys
might have discussed before but I didn't found it in past mail
archives,
Some of the objections against having new components were along
those lines (i.e. "Why not make modules?").
The problem is not with modules (I quite pushed for modularization
in "Commons RNG" and "Commons Numbers"), it is with "Commons Math"
requiring so much work to tackle the backlog (some identified issues
are _years_ old), in addition to the work for modularizing it.
I don't think that it is acceptable to release code within a new suit
("module") without fixing it too.
And other people here indicated that no Commons Math release should
remove any code for which no alternative exists.
So, "Commons Math" is stuck.
Gilles
though I saw a fork of CM made last year and "that" code base is
doing
exactly what my doubt is. i.e sub-component as maven module.
Regards,
Amey
On Tue, Aug 15, 2017 at 7:56 PM, Gilles
<gil...@harfang.homelinux.org>
wrote:
Hello.
[Time for a new episode in our "Ripping CM" series.]
How about creating "Commons Geometry"?
The rationale is comprised of the usual suspects:
* Smaller and more focused component, hence:
- Consistent development and maintenance.
- Consistent release schedule, not encumbered by
changes (and endless discussions) in _totally_
unrelated code.
- Potential for attracting contributors not
interested in maintaining the (growing) backlog
of CM.
* Self-contained: 96.3% of the "o.a.c.math4.geometry"
package have no dependency except:
- 4 classes now in "Commons Numbers".
- 2 methods and 1 constant in "MathUtils".
- CM exceptions. [Creating alternatives for those
will probably be the most time-consuming part of
the porting work.]
Moreover, none of the code in the "o.a.c.math4.geometry"
package is used by another package of CM.
A new component would give the "geometry" codes a much
better chance of being (confidently[1]) released, since
CM is "stuck" for the foreseeable future.[2]
WDYT?
Gilles
[1] There seems to be only one issue reported in JIRA
that pertains to "geometry".
[2] 54 issues yet to be fixed before the 4.0 release;
which, at the current rate, would lead to after 2025
(a very rough guess, I admit).
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
