Why so few Python jobs? (and licenses)

[Paul Rubin]

Cliff Wells <logiplexsoftware at earthlink.net> writes:
> I have to admit I hadn't thought of that.... :( Still, to be honest,
> I consider GPL'd/open source stuff to be a matter of goodwill, and
> if I submitted a patch to a GPL'd author that means that I am
> obviously using his code to a significant degree and the patch is
> just my small thanks for that.  If the author later profits from
> that it's all the better, as far as I'm concerned, as this will
> undoubtedly encourage him/her to develop more code (and maybe even
> let him eat, if he needs to).  People who resent that sort of thing
> are perhaps missing the larger picture.

I mostly agree with this sentiment--it does depend somewhat on the
situation though.  If I correct some typos in the manual at my
leisure, sure, that's a gesture of goodwill.  If I spend $5000 worth
of development time writing and releasing a feature for GIMP that I
could have had instead by paying $500 for a retail copy of Photoshop,
the GIMP author has (at least monetarily) benefitted from my effort
more than I've benefitted from his.  (Note: that GIMP example is made
up, not something that really happened to me).

  Most (admittedly not all - large projects like Linux being an
> exception) GPL'd projects 90% the result of one or two people's
> efforts with small improvements and fixes being submitted by others

I guess that might be true--I'd be interested in seeing a real survey,
based on, say, examining the CVS histories of projects on SourceForge. 

> (who, as I said earlier, are already benefitting from the existence
> of that project - probably more than enough to compensate for their
> small contribution).  

OK, though one should not underestimate the value of "small"
contributions, especially considering that some may have been made at
"gunpoint".  E.g. I mentioned in another post I once spent a frantic
all-nighter finding an Apache module bug that was crashing my
employer's web site.  The result was a two-line fix (it was a malloc
error).  If I did emergency repairs of someone else's software for a
paying client, I'd charge at least 2x-3x my normal consulting rate per
hour.  By that reckoning, the two-line fix was worth enough to my
employer to pay for several copies of a commercial server, because of
the time spent in a critical situation.  (Of course the commercial
server would have had its own bugs...)

> I suppose those same people resent the FSF selling copies of GPL'd
> books.

I hope not--a main point of the GPL is you're ALLOWED to sell copies
of GPL'd stuff, as long as the recipient isn't restricted from making
further copies.  The FSF has in fact sold copies of a book I wrote and
I don't object at all.  It annoys me more if I write a program and a
commercial publisher publishes a proprietary book about it--I'd rather
that people buy the FSF book.

> I think licensing GPL'd code is a good idea if it helps support the
> efforts of open source programmers.  

I guess that can be true in some situations.  Other times it's the
opposite.  GCC has benefitted tremendously from contributions of
backends for various machines, front ends for alternative languages,
etc.  If proprietary versions were allowed, those improvements never
would have been released.  So I think more free Python improvements
would result if Python were GPL'd, though I'm also ok with its present
state of licensing.

Anyway, the main thing IMO is not whether something is GPL'd, but that
whatever the author decides to do, they announce their choice up front
and don't spring surprises on contributors afterwards.

> Getting paid to develop open source is the best of both worlds and
> benefits everyone.

Yes, that's a very good feeling :-).

> Very true, I was thinking more along the lines of _who_ you choose
> to give your money to.  Since I use open source/GPL stuff all the
> time, if I can justify giving them money, that's what I prefer to
> do.  That's why I buy the RedHat CD's rather than just downloading
> them.

I like Red Hat--they've written a lot of good code, and after
stumbling a little in their early days, they've been consistent about
GPL'ing all of it.  I think it was Eric Troan who described it with
some surprise--they found the more they gave stuff away, the more
successful they were.  (However, since I'm a cheapskate with currently
low income, and since Redhat CD's contain a fair amount of code that I
wrote, I feel like I've already contributed enough to them, so I don't
go out of my way to buy the CD's--but I'm glad that you do <g>).