Hello.
On Fri, 22 Dec 2017 17:00:10 -0700, Gary Gregory wrote:
Hi All,
In light of the [math] work being done, does anyone have thoughts on
what
should happen to the stat packages? As in stay in [math] or [split]
into a
new component or module?
Gary
It will come as no surprise that, in principle, I favour "split".
[It is telling that the initial proposal[1] for the "component"
listed "statistical algorithms" and "mathematical algorithms" on
a par (with the former cited first!).]
There are only a handful uses of classes from "o.a.c.m.stat" by
other parts of CM, most of which in "o.a.c.m.distribution" and
the remainder in "o.a.c.m.ml.clustering".
Setting up of "Commons Stat" would not be an easy one like "Commons
Geometry" since package "o.a.c.m.stat" depends on classes in
* o.a.c.m.linear
* o.a.c.m.analysis
* o.a.c.m.distribution
Package "linear" itself requires a lot of work (but we could perhaps
hide its API from the POV of "Commons Stat" users).
Also to be noted: some design issues were uncovered in "stat" (see
JIRA).
I think that the contents of "o.a.c.m.distribution" would nicely fit
as a module of "Commons Stat". [Quite some work has been done on
"distribution" since the last release of CM, thus we could imagine
to release versions 0.x of "Commons Stat" with contents gradually
added as issues are fixed.]
Gilles
[1] http://commons.apache.org/proper/commons-math/proposal.html
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
