commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Phil Steitz <>
Subject Hipparchus fork of Commons Math
Date Wed, 01 Jun 2016 13:21:43 GMT
A couple of months ago, some contributors to Commons Math decided to create a new project,
starting with a fork of the [math] sources. This was a hard decision for us to make.  It is
not the intent of this message to rationalize our decision or to in any way disparage Apache
Commons Math.  I will just try to make clear our reasons for starting our new project, Hipparchus.
 I can of course only speak for myself and take accountability for my own actions.

There are three things that led us to start (privately) discussing the idea of a fork.

0) [math] has become too large in our opinion to stay in Commons.  We had talked previously
about moving to TLP; but our assessment (see 2) below) was that we were not large enough or
well enough to establish an ASF PMC at the current time and we did not think a trip through
the Incubator would help.
1) The versioning rules in Commons were too restrictive for us.  We respect the PMC consensus
that breaks can only occur with major version changes and package name changes, but we have
had a very hard time with that in [math].
2) The [math] community was having a hard time reaching consensus on code-related decisions.

We decided that starting an independent project based on a fork was something that was worth
trying.  We did this "in private" partly because we did not want to generate negative discussion
or sentiment here and partly because we just needed a "cooling off" period.

If there is blame to be laid for 2) above or for the decision to fork, you can place that
blame on me.  I know that I was often - and will often be - wrong in my positions and I know
that I wrote some posts that I am not proud of.  The archives are public, so people can come
to their own conclusions.  Please don't read anything more into this message than the simple
facts above. The issues in 2) can only be understood fully by reading the archives and digging
into the code.  They are not just "change" vs "no change" - they are about what kinds of change
are positive and balancing user and developer needs.  We do better @apache when we talk about
actually doing things, rather than talking about talking about things; so I encourage all
involved to avoid making generalizations about Commons or [math] or any of the individuals
involved and instead to go find some itches and scratch them.

Over the 13+ years that I was part of the Commons community, I saw many contributors come
and go and most components turn over their entire development teams multiple times. The [math]
component is no different.  There have been over 100 different contributors and new ones are
showing up all the time. It was a little odd, frankly, to have some of the earliest contributors
remain active so long.  It is natural for the community to move on and I am sure it will.


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message