[Tfug] GPL Worthless?

[Bexley Hall]

Hi Zack,

--- On Sat, 9/8/12, Zack Williams <zdwzdw at gmail.com> wrote:

> To understand the GPL, you have to understand the FSF and Richard
> Stallman.  Here's a guy who, working at MIT, got frustrated that code
> was being taken and put into things like printers, the result of which
> was that they were buggy and impossible to fix or tinker with, so he
> came up with a license-based solution to it.

I'm well aware of Stallman and the FSF (he was a grad student when
I was an undergrad).  I suspect he would be even *more* outspoken
on the "problem" (flaw?) that the GPL can't recover from than *I*
have been (here).

It is unfortunate that he was unable then (or since?) to realize
how pervasive writeable control stores, firmware/slushware, etc.
would become (and *have* become) to try to counter the effective
threat they pose against the ability to modify the "work"/product
in question.

At the AI Lab, he surely was exposed to this sort of technology
(*we* were as undergraduates so hard to imagine him NOT!).  And,
regardless, in the decades since then, even a blind man could
see how commonplace this sort of implementation has become!

> The GPL put the desires of savvy coder/users (the only kind at the
> time it was written) ahead of 3rd party developers, who wouldn't
> release source and thus their products couldn't easily be tinkered
> with or fixed.   And he did it by writing a license that prevented any
> of his code from being included into distributed products without
> being also shared to the recipient. In the end, he hoped to be able to
> run a completely free operating system, and in this goal he was
> successful (see here: http://richard.stallman.usesthis.com) The rest
> of the issues (releases being buggy/unbuildable, spaghetti code, etc.)
> are implementation details that would be present with any software
> product or release.

Note that Athena was contemporaneous with RMS and didn't feel inclined
to impose copyleft on its sources (hence the "MIT License").  I suspect
that was the largest "shared" software development in history, at that
time.
 
> Your goals are probably different, on some level.  You might care much
> more about a specific job being performed than forcing the hand of
> others who want to improve your work.   You might be in the same
> situation as RMS with regards to your household appliances, but with
> little power to effect change.

My point goes to the *effectiveness* of the GPL vs. non-copylefted
(FOSS) licenses.  If I can bury much of a product/work's functionality
*in* something "closed", *legally*, then what value does the GPL
offer over a BSD/Apache/MIT/etc. style license?  I.e., the *spirit*
of the GPL hopes to encourage me to "share" -- but the *terms* of
the license don't require it (given the number of ways I can workaround
its restraints).
 
> Everyone violates software licenses - I love this bit from a Microsoft
> slide deck, where their biggest competitor by far is "unlicensed"
> Windows:
> 
> http://www.osnews.com/story/21035/Ballmer_Linux_Bigger_Competitor_than_Apple
> 
> If enforcement of licensing was complete and total, I'd wager we'd
> have a lot more "free" (by any meaning of the word) software use out
> there, just to avoid the entire licensing hassle.

As a "writer/publisher/designer", the whole licensing/patent issue
is a non-starter (for me).  I'd rather spend the effort building a 
better mousetrap than *litigating* against someone who has "stolen"
from me (that doesn't mean I want to encourage theft!).

This is especially true with products that are dynamic in nature!
It's not like selling cars and whatever you bought today is what
you're stuck with tomorrow.  If I release an upgrade, you *gain*
value if you've purchased my device.  If that upgrade doesn't
work on a counterfeit (or otherwise copied) product, you lose!

And, try calling with a support question -- "Sorry, you'll have
to contact the folks in China that built the device that *you*
since we have no record of your purchase from *us*!"

On the other hand, I recognize that other commercial concerns
are in business to make money.  So, if I want them to adopt my 
designs (which is the only way they will evolve beyond my
implementations), then I need to have terms in place that allow
them to *safely* hide their improvements -- without ever wondering
if I'm going to litigate at some future date.  Or, have to do a
clean room reimplementation of some core technology that I've
developed.
 
> We haven't even gotten into the patent system nightmares that clauses
> of the Apache and similar licenses try to protect from, while allowing
> commercial use like the BSD licenses.

The same hassles and loopholes apply to the patent system (I've
spent a fair bit of effort, over the years, avoiding "independant
claims" in order to skirt infringement).  And, if you've ever spoken
to a patent *examiner*, you'd realize just what a *game* this
whole process is!  :-/

When we filed for my first patent, software was so novel a field
that the whole issue of patentability was actively being argued.
We sidestepped it by describing our algorithms as bits of *hardware*
(which is clearly patentable!).  I can't count the number of times
the phrase "in one embodiment" appeared in the initial filing!
(i.e., intended as "in a SOFTWARE embodiment"!  :> )

Most "good" (?) engineers look at patents and say, "Sure!  That's
OBVIOUS!" and, as such, don't even think about patentability.
(some, OTOH, like *counting* patents as if it was a currency of
sorts)  I'd prefer to *do* things with the time that would otherwise
be spent arguing with patent atty's, etc.  (though that's because
I'm a lousy businessman who finds more joy in the Engineering than
the Sales  :-/ )

[I designed a little gizmo to hold *screws* in place.  "Patent
this?  Are you crazy?  Haven't you ever heard of tweezers??"]

[BTW, Zack, let me know when you've got some time -- I'd like
to take a peek at that gizmo.  FWIW, it *truly* seems to be
"open"]

--don