[Tfug] GPL Worthless?

[Bexley Hall]

Hi John,

--- On Sat, 9/8/12, John Gruenenfelder <jetpackjohn at gmail.com> wrote:

> Richie is correct that the FSF is one of the main bodies
> that pursues GPL violations.  Of course, they don't do
> this proactively.  The actual copyright holder has to bring
> it to their attention and in most cases litigation never
> occurs because the guilty party will make some effort at
> releasing source (and remember that you need only release
> your changes; you don't have to do anything for just
> *using*).  The FSF doesn't just troll the Internet for
> possible violations.

I'm not claiming they should!  But, I see it only effective 
with folks who *embrace* its ideals.  Like locks on doors
keeping honest people honest, etc.  ("10 item or less"
signs on express checkouts, etc.)

It's just too easy to circumvent the spirit of the license
while complying fully with its *terms*.  And, for those
who wish to do neither, even *that* is very practical
nowadays in a world where a "Chinese garage shop" can
flood a market with inexpensive product in the time that
it takes anyone to even *notice* that they've not complied.
Then, swap two of the pictographs and start all over
(with the same *product*!) again.

[This was a common problem in the 80's where Japanese
manufacturers would brazenly counterfeit US-made arcade
games then "go out of business" before any legal actions
could be taken against them -- only to crank out the
*next* counterfeit immediately thereafter!]
 
> I can't think of any offhand, but if you follow
> Slashdot you will see stories about GPL violations and
> resolutions every now and then.
> Also, I would argue that the commercial world is not the
> be all end all for copyleft use.  The honor system does, in
> fact, exist and does work among many FOSS developers and
> proponents.  The fact that I *choose* to honor the terms of
> the GPL when applicable means that the license does indeed
> have worth and power.

But that spirit has been alive and well in academia for
decades.  And, there is nothing in a (FOSS) non-copyleft license
that precludes it.  E.g., you can take an Apache-licensed
work, modify it *and* publish your changes under the same
license.

I see the real value of any "open" licensing scheme to be
*exactly* in the commercial arena.  That;s where the vast majority
of people "consume" their software -- and, where most software
*exists*!

(is the BIOS for your desktop machine *open* for you to inspect
or modify?  will your PC run without it?  what about the code
that makes your mouse work?  or keyboard?  or CD/DVD drive?
or disk drive?  or GPU? -- and we haven't even left your desktop
to explore your stove, refrigerator, microwave oven, any of the
dozen processors in your automobile, STB, furnace, BT earpiece,
the remote control for your TV, etc.)

I doubt many people would be keen on replacing the code in
their ECU (engine control unit) with something "open" and
heading off on a 100 mile trek to feenigs.  Regardless of
how well the FOSS community *claims* to have tested it!
Heck, I doubt they'd be willing to drive down *Grant*
without a laptop sitting on the seat ready to reload the
original ECU image!

[I designed an electronic ignition system for one of my
vehicles as a teenager.  You can *bet* I left the original
coil, points, etc. in place so I could QUICKLY remove *my*
electronics and restore the car to its original configuration!
Don't want to find myself stranded while on a date, etc.]

As the popularity of the desktop wanes, I think this becomes
even more significant.  As more and more "appliances" (even
traditional appliances -- stoves, refrigerators, etc.) around
us begin embodying more sophisticated control algorithms and
connectivity features, you're going to really *want* to
pressure those suppliers to comply with the *spirit* of the
license.  Or, not at all!

E.g., our washing machine and drier have an annunciator that
chimes when the load is finished (or to announce other conditions
that require intervention).  It *also* serves as the function of
a "keyclick" (chirp, chirp).  You can disable it *entirely*
(meaning no "notifications" at all!) but can't *just* disable
the keyclick (why do I need an annoying little chirp to tell me
that the indicator lights in front of my eyes have just changed?).

This has got to be a one line change to the existing code, someplace.
Comment out the statement that turns the chirp *on* when a keypress
event is detected (you can leave the statement that turns it *off*
100ms later as that won't typically be doing anything).

Or, maybe I'd like to be able to start the washer halfway through
the wash cycle ("short wash").  etc.

Perhaps I would like to modify my telephone answering machine
so that it looks at the CID before answering and, for blacklisted
(or non-whitelisted) numbers, simply refuses to answer.

Surely these types of changes are no less desirable than adding
support for a new basic data type to a DBMS; or, CGI support
to a web server; or client side caching to a name resolver; or,
support for some new fangled I/O device to a kernel...
 
> Personally, I always use the GPL unless their are some
> conflicting circumstances.  I use it because I like what it
> does and what it represents.  That said, I rarely do
> anything related to hardware and you make a very good
> argument for using a BSD-like license in that case.

But software is increasingly moving onto platforms that
can comply with GPL in spirit -- yet offer you *none* of
the benefits that GPL is intended to convey!

Every processor in a design can run closed firmware and the
system itself can still be in compliance.  Nowadays, adding
processors is almost /de rigeur/!  And, with all the device's
functionality embodied in that firmware, there is little
you can do to fix/workaround/alter the product's performance
with regard to that subsystem(s).

I can design a system, today, that puts an abstraction layer
*under* the linux kernel making it look like *most* of a
conventional "PC".  I can then run a stock kernel.  And,
never give you *any* of the details about the implementation
of that abstraction layer (think: virtualization).  And,
I am legally safe!

But, any bugs in that virtualization (or, features that I
decided NOT to implement) will remain that way regardless
of your desire or efforts to fix or improve upon them.  And,
the "Runs Linux" that I *use* to appeal to folks who think
they'll be able to tinker with it actually increases my
profitability by *screwing* more purchasers drawn to that
"flame".

Those folks would be better off seeing my device as a "closed"
system and *avoiding* it -- than duped into thinking that the
GPL has some *value* in making my system more "supportable".

<frown>

Crappy idea to start the day with!  :-/  Maybe I'll go pull some
weeds in the yard...

--don