[Openmcl-devel] user submissions and open-sourceness

[Erik Huelsmann]

On Mon, Jun 16, 2008 at 10:34 PM, Hamilton Link <helink at sandia.gov> wrote:
> I'm not affiliated with Clozure.  I've worked with mcl/openmcl/ccl for most
> of ten years, have spent modest amounts of time scrounging around in the
> code, have read all the implementation docs a couple times over, and such.
>  I have submitted bug fixes to a few things (tcp streams iirc, some GUI
> stuff... maybe more).  I implemented the cocoa inspector and the OpenGL demo
> and tested out the early hemlock stuff a fair amount, and now and then I've
> helped out rsynching the server or running this or that CLTL2 regression
> test suite.
>
> I'd pretty much never try porting ccl to a new platform or debugging a brand
> new port.
>
> I think one reason ccl works as well as it does is because although it is
> open source, there are extremely knowledgeable, dedicated people who spend a
> lot of time maintaining and supporting it.  SVN, Apache, and many other
> "open source" projects have a similar model dividing the world into gurus
> and non-gurus, and I think it works well because users get a quality product
> and as new gurus arise they are generally welcomed... and when non-gurus
> suggest a patch it gets sanity checked and incorporated in an orderly
> fashion.
>
> My point is... I don't think anyone can reasonably expect useada.u to jump
> in and work on a 32-bit x86 port (or <user> to jump in and work on <major
> rewrite>).  It's silly to suggest it.

Why? Where do you think Apache or Subversion gets its new gurus?
Exactly by doing what I did: encourage people to contribute, coaching
them in their efforts.  The hardest of problems are often addressed
(we find in Subversion) by people who have a fresh look at an old
problem.

>From what I see happening in SBCL is that - even for software which
has far less documented sources than Subversion - it's the newcomers
who do the ports to the new platforms: those are the ones with the
itch...

>  But I would also encourage everyone
> to contribute to ccl in whatever capacity they feel able.  Such help has
> always been welcomed, whether it has been running their application in
> multiple versions of ccl, porting and testing lisp libraries, reporting bugs
> and problems while developing in lisp, helping debug or submitting probable
> fixes, updating or editing documentation, or developing new functionality
> for ccl.

Well, at least that sounds a lot different from what the OP claimed. I
must say I find it a lot more encouraging to anybody considering a
contribution!

Thanks for your views!


Bye,

Erik.