Hello.
Referring to
https://issues.apache.org/jira/browse/MATH-1346
The class mostly adds a trivial layer (for a subset of the
distributions available in CM) around "RandomDataGenerator"
(which itself was a trivial layer around the sampling API
directly provided by the distribution classes).
"RandomDataGenerator" will disappear, the non-trivial
functionality being replaced by "RandomUtils" (cf. MATH-1341).
Hence, "ValueServer", being the sole remaining class in CM
that depend on "RandomDataGenerator" would need to be modified.
However, AFAICT, functionality in "ValueServer" is either
duplicating similar functionality available elsewhere in CM
(e.g. "fill" vs "sample" in "AbstractRealDistribution") or some
trivial task that should not be bothered with by CM (e.g. read
a file line by line and convert its content to a "double"[1]).
Did I miss some unique contribution of this code?
IOW, can someone come up with a case that would require a
"ValueServer" object in new applications in preference to the
alternative, direct, way for getting the same functionality?
If not, I propose to deprecate this class.
[As usual, old applications can always keep a dependency to
CM v3.x if they use "ValueServer".]
Regards,
Gilles
[1] Obviously, this has nothing to do with random data
generation.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
