Subject: Re: Successful FSBs
From: "Tim O'Reilly" <>
Date: Mon, 04 Nov 2002 18:19:29 -0800

On 11/3/02 9:59 PM, "Stephen J. Turnbull" <> wrote:

> How about Aladdin Ghostscript or RealSSH or Windows Qt or libdbN, N =
> most recent?[1]  You can't just "take" them if you're going to sell the
> result, but you can just take the source and fix it, then buy the
> commercial license.[2]

> [2]  I doubt Tim has any more problem with this than I do, but it is
> not OSD open source, nor free software.

No.  No problem at all.   "Source available" but proprietary has many
benefits.  Heck, *every* licensing philosophy has benefits.  The trick is to
figure out which is best for your project, and why.

See my Nature piece on scientific publishing, which I've mentioned before on
this list.  (Can't give a URL now, since I'm on a plane, but go to if you can't find it in google with something
like "Tim O'Reilly information wants to be valuable".)  That's the piece
about what Bill Gates and Larry Wall have in common.

>   eric> if there's something I don't like about them, I can't just
>   eric> fix it myself and contribute it back to the maintainer --
> Good for you.  But let me tell you a story.
> I've recently gotten a bunch of bug reports about XEmacs from Jamie
> Zawinski, in a couple particular about the event loop.  There is only
> one person in the world who might know more about the XEmacs event
> loop than Jamie, and that person has no visible net presence, doesn't
> answer mail, etc.  Guess what?  Jamie hasn't contributed a fix, and I
> don't expect him to do so.  Nor do I think there's any reason why he
> should.  He retired from XEmacs maintenance 8 years ago, maybe more.
> Jamie doesn't even have a copy of the XEmacs sources, it seems---he
> uses Red Hat's RPM.
> Man, if _jwz_ doesn't build his own, let alone contribute bug fixes to
> software he designed and built from the ground up[3], how are you
> going to convince a busy programmer that he should do as _you_ do?

Great story.  I think that the number of people who actually contribute to a
given piece of code is quite small.  Which is why I keep harping on the
importance of open source architectures and things like cpan.  A much larger
community can contribute to the success of Perl by writing a module and
contributing it to cpan than can contribute by rewriting some piece of the

> The point is that we have good reasons, good enough for us, for using
> open source.  Others have (as far as they are concerned) better
> reasons for _not_ using it.  If we're going to convince them, we need
> to address _their_ problems, not project _our_ problems on to them.

Very important point!

Tim O'Reilly @ O'Reilly & Associates, Inc.
1005 Gravenstein Highway North, Sebastopol, CA 95472