incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Burrell Donkin <>
Subject Re: UIMA [WAS Re: Suspending Projects]
Date Fri, 20 Feb 2009 11:14:04 GMT
On 2/20/09, Janne Jalkanen <> wrote:
>> My pet hypothesis (or maybe I'm just looking for excuses) is
>> this: UIMA is heavily used in academia.  Now academics have no
>> problems with open source, to the contrary.  But they have an
>> overwhelming need to publish and build up a reputation.  So
>> they like to publish their source code on their own web site,
>> where it's clear it's their work, rather than contribute to
>> some community effort.  If you look around, you'll see all
>> manner of university efforts around UIMA, but very little of
>> that code finds its way back into the ASF repo.
> FWIW, I've noticed this with JSPWiki too - there are at least three
> (maybe four-five) separate forks of the JSPWiki codebase within the
> academia, with little contribution back to us.
> There are a few additional reasons for this:
> * The authors feel that their code is not mature enough for inclusion,
>   and often they may be right - code is created to solve a specific
>   problem and on a proof-of-concept level with little regard to
>   e.g. maintainability.
> * Often, the forks are not maintained after the paper is completed and
>   published.  The authors move to different subjects, and have little
>   interest to keep working on the codebase (which might be needed in
>   case of invasive modifications)

I know that we usually try to strongly encourage companies not to use
apache as dumping ground but I wonder sometimes whether it might be
useful to accept more contributions of proof-of-concept code
especially from academia. It's often easier to start from some proof-
of-concept code than from scratch.

> * It brings them no benefit to contribute back, since there is little
>   attachment to the project itself.  What's important is their
>   modifications; the project itself is just necessary noise which
>   allows them to tinker with their grand ideas.

I wonder whether we could do more on our sites. We often do Powers
style linking (for OpenJPA powers WebSphere) which link into wider
FOSS and commercial communities but doesn't seem particularly
appropriate for research.


> (Yeah, I worked for the uni for three years ;-)
> /Janne
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

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

View raw message