[
https://issues.apache.org/jira/browse/NUMBERS-148?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17162419#comment-17162419
]
Mohammad Rezaei commented on NUMBERS-148:
-----------------------------------------
bq. from the perspective of someone trying to understand the solver, the
scaling functions are just added noise.
There is more than one way to reduce noise. Putting things in a method is one
way employed in the code.
bq. It could as well be seen as the user's responsibility to "prepare" the
inputs
Simplex solvers are old -- invented in 1946 before hash tables or binary trees!
40 year old implementations written in Fortran have scalers.
bq. few people maintain it ... is IMHO a sad example of this issue.
I get where you're coming from and I think it's great to worry about the health
of the library, but that should be all the more reason to keep the public API
to a minimum, not expand it when there is no demonstrated public need. I mean
that very seriously. I look around the library and I don't see a coherent
offering, just a bunch of of classes that seem to do random stuff, the need for
which is exceptionally situational. I wouldn't even know to look for them if
they weren't famous (like Simplex, or Newton-Raphson, which I tried to use
recently but gave up on).
I guess the move to break up the library is the long term solution for the
issue. I would suggest that if/when the time comes to create a
commons-linear-programming, that would be the time to decide how to reorganize
the public API and Simplex implementation.
> Extract double exponent manipulation routines from commons-math to numbers
> --------------------------------------------------------------------------
>
> Key: NUMBERS-148
> URL: https://issues.apache.org/jira/browse/NUMBERS-148
> Project: Commons Numbers
> Issue Type: Improvement
> Components: core
> Reporter: Mohammad Rezaei
> Priority: Minor
> Time Spent: 20m
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
