[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] is condor open source?
- Date: Wed, 11 Apr 2007 10:21:59 -0700
- From: Jonathan Day <jday@xxxxxxxxxxxxxx>
- Subject: Re: [Condor-users] is condor open source?
On Wed, 11 Apr 2007 12:16:16 -0400
"Nick Coleman" <nacoleman@xxxxxxxxx> wrote:
> On 4/11/07, Alex Brown <Alexander_Brown@xxxxxxx> wrote:
> > Glad to hear this. Regarding what Nick Coleman said about the
> > developers' time I personally think you might be able to spend less
> > of it on support if source were released.
> In the long term yes, in the short term no. That's why (as Dan said)
> they are working towards open source, but that takes time. In the
> meantime there are still all of the short term concerns on the table,
> which would have to take a back seat.
> I'm whole-heartedly in favor of Condor being open source, and of open
> source development in general. But I think it's a common myth that
> open sourcing a project magically fixes all (or even most) of its
> problems. I think it works to everyone's favor to have security
> issues hammered out to the Condor team's satisfaction before the
> source is release to the general public.
When Netscape's browser (a relatively small project) was Open Sourced,
it was hailed as a major step forward. It was then completely
re-written as the original code proved impossible to develop under such
a model at the level of monolithic complexity it had reached.
That experience alone should have taught people that Open Source is not
a solution, merely a mechanism by which a solution can be reached.
I'm a passionate advocate of Open Source and have a strong dislike of
"Secret Source" (projects nominally Open Source usually for funding or
political reasons, but which are hidden in the dark alleys of the
Internet), but I would point out two things:
Firstly, by my understanding, Condor isn't remotely Secret Source. It's
well-known and doesn't seem to have any problems with keeping it that
way. The source isn't trivial to get, but have you seen the
requirements for some NASA projects? Some of their CFD projects require
signed and mailed disclaimers on corporate notepaper with proof of US
citizenship and an extensive reason, with copies faxed to the project
manager. That's to get any code at all. By comparison, Condor is
laid-back, relaxed and chilling out by the pool.
Secondly, it would be a grave error for Condor to repeat the Netscape
experience. I couldn't even guess at what would be the right way for
the development team to move forward, but en-masse code dumps have
generally not been useful to people.
(Netscape's not the only one to make that mistake. Of the many hundreds
of Open Source code dumps and complete project code drops by IBM and SGI
over the years, how many have they found themselves capable of moving
forward with? How many Open Source developers picked up SGI's OB1
security code, or IBM's DAISY live opcode translation software? Know
anyone using HP's plug-in scheduler system for Linux, or the complete
port of STREAMS?)
If the Condor team thinks that spending another few years on code
review, cleanup and maybe segmentation will lead to a better, more
productive, release under an Open Source license, then they deserve
credit for being truly aware and support for attempting what is