🚀 go-pugleaf

In article <ki1jj3INN68v@early-bird.think.com> barmar@think.com (Barry Margolin) writes:
>In article <1991Nov11.221754.16254@alphalpha.com> nazgul@alphalpha.com (Kee Hinckley) writes:
>>I've found very few "end-user" apps that had bugs I couldn't work around.
>
>Sure, you *can* work around most bugs.  Does that mean that you should be
>*forced* to?  Why should vendors have so much control over their users?

There is no difference at all between being "forced" to work around a bug
and being "forced" to modify the software to a non-programmer.  Frankly I
usually find it much faster to find a work-around to most GNU bugs than
to try and fix the software.  You're use the word "force" implies that
I would prefer to modify the source code - I wouldn't.  I'd prefer that
the vendor fix the bug.

>Who should be in control here, us or them?

Them if you believe that software should be a market-driven product.
Us if you you believe that software should be driven by communist principles.

>What if your bicycle dealer didn't allow you to perform your own
>adjustments and repairs (I'm not going to repeat the car analogy, which you
>shot down pretty well)?

There's very little point in making analogies between software and physical
objects.  If a bike manufacturor somehow managed to keep me from getting to
the internals then clearly I'd have to make a decision as to whether the
features of the bike made it useful enough that I wasn't concerned about
having a sole source for repairs.  That's the same decision I make for
commercial software.  I should note that car manufacturors in fact do do
this.  They just also happen to allow a second tier of authorized support
agencies who have more internal access to the car than I do.  There's no
particular reason that a software company couldn't do the same.  In fact
many companies do license their source to VARs who then make customer
specified changes.  I don't see anything wrong with this model.

>Well, we have over 200 Macs, so an upgrade could cost us as much as $10K.
>It would be much more cost-effective for us to pay someone for a few hours
>work.  And it would be even more cost-effective for us to let one of our
>systems programmers fix it.

There is a misconception here.  It's the assumption that consultants
are consultants are consultants.  That isn't true.  You can't just hire
someone to come in and fix a few bugs in a few hours.  Software is *not*
like an automobile.  Cars tend to be several orders of magnitude more similar
than different software packages are.  There are thousands and thousands
of commercial software packages available.  If suddenly the source code
to all of them were available then yes, you could probably get bug fixes
to some of the major ones - since it would be cost effective for a consultant
to specialize in those.  But for the vast majority of them it would not
be that easy.  We spent a man-month on g++ trying to fix a single multiple
inheritance bug, and we talked at length to Cygnus and others about it.
As far as I know it's still not fixed, although we reported the problem
a year ago.  (I'll grant, if we could afford to pay Cygnus it might be
fixed by now, but we can't.)  For the vast majority of programs there
is only one group of people who are going to be able to quickly and
efficiently fix bugs - and those are the people who wrote the program.
For them it is *far* more efficient to charge for the program and provide
the fixes at very low cost.  Giving away the program and charging for
support requires much more overhead (contracts, time between fixes,
unreliable income stream, issues of how to spread the cost over
many users when only a few are paying for support...).  It just doesn't
make economic sense.

But then, I don't think the reasoning behind the GPL has anything to
do with economics.  I think it's a moral issue.  Information is becoming
more and more important.  Treating information as a commodity creates a new
group of disenfranchised people.  If you believe that software is
"information" in this sense, then the GPL is a morally correct solution.
I think there are two flaws to this argument.  I don't believe that
software is information (although it may make use of it).  And it is
not clear to me that global enforcement of the GPL would actually serve
the greater public good; I'm concerned that the drawbacks might outweigh
the benefits.

That said, there's no reason not to use the GPL if you believe in it.
It's just another copyright.  In my mind it's too restrictive.  When
I release code I do so under the MIT X copyright.  I have no problem
with people using it in commercial software, I just ask (but don't
require) that they feed back any changes to me.  I take that path
because I believe that in general I can trust people to do the right
thing.  The GPL assumes that you can't.


-- 
Alfalfa Software, Inc.          |       Poste:  The EMail for Unix
nazgul@alfalfa.com              |       Send Anything... Anywhere
617/497-2922                    |       info@alfalfa.com

I'm not sure which upsets me more: that people are so unwilling to accept
responsibility for their own actions, or that they are so eager to regulate
everyone else's.