wxPython Licence vs GPL

[Mike Meyer]

Those who can not afford the software are excluded for that end
product even though they may have worked on the source where 99,99% of
the restricted licensed software is based on.

Well, they chose to make it available to others for reuse. But
software "unavailable to those who can't afford it" is better than "no
software at all"

However I make a poor defender for the GPL because, as you can read in
my previous posts, I don't really believe in it.

The question is wether or not it believes in you

I believe in GPL'ed software - I use it regularly. On the other hand,
I don't believe that it represents the best license to release
software if the goal is to improve the lot of humanity. The
restrictions are on "distribution", not on use, so it doesn't really
keep people from using said software commercially. For instance, one
or more of your examples may have been worth developing for internal
use. They then decided there was a profit to be made in distributing
it commercially, and proceeded to do so because they could. Without
the profit motive, they may not have done the extra work involved in
preparing the IP for distribution and doing the distribution.

Personally, I release stuff under a BSD-like license, historically
having included requirements that I be notified of bug fixes, and/or
that I be given copies of commercial software that included my code. I
eventually gave up on them as unenforceable.

<mike

 

[Martin P. Hellwig]

Mike Meyer wrote:

Well, they chose to make it available to others for reuse. But
software "unavailable to those who can't afford it" is better than "no
software at all"

That I do not agree with, I think it depends on which your side of the
fence you are.

For instance I have a specific problem, there are currently 2 product
available that come close to solving it, one costs $24,999 and the other
is above that. That is about $23,999 above what I can afford to solve my
problem, so I have the option to leave the problem as it is or try to
tackle it myself.

Stubborn that I am, I am currently creating my own solution, knowing
well that other solutions exist and I can only make a poor copy of those
already existing effort

The question is wether or not it believes in you

I believe in GPL'ed software - I use it regularly. On the other hand,
I don't believe that it represents the best license to release
software if the goal is to improve the lot of humanity. The
restrictions are on "distribution", not on use, so it doesn't really
keep people from using said software commercially. For instance, one
or more of your examples may have been worth developing for internal
use. They then decided there was a profit to be made in distributing
it commercially, and proceeded to do so because they could. Without
the profit motive, they may not have done the extra work involved in
preparing the IP for distribution and doing the distribution.

Yeah well, GPL works reasonable well but perhaps not for what it was
intended.

Personally, I release stuff under a BSD-like license, historically
having included requirements that I be notified of bug fixes, and/or
that I be given copies of commercial software that included my code. I
eventually gave up on them as unenforceable.

Thats the trouble with restrictions, how do you enforce them, with
license I don't found it worth the hazzle. BSD/MIT style license is a
good substitute of no license at all.

 

[Mike Meyer]

Mike Meyer wrote:

That I do not agree with, I think it depends on which your side of the
fence you are.

Actually, I think I underspecified the statement. "Better for who" is
the question. No software is the same as software you can't afford. On
the other hand, so long as the price is lower than the cost of
recreating the software for someone, then it's better for society as a
whole if it exists at all.

<mike

 

[Paul Rubin]

On the other hand, so long as the price is lower than the cost of
recreating the software for someone, then it's better for society as
a whole if it exists at all.

I don't think that's correct. Having nothing can be better than
having something, because then there's an opening to fill. The price
isn't even an issue: it happens even with free programs. If someone
needs an XYZ program and none exists, maybe they'll write a good one
and release it. But if a bad XYZ program already exists, they'll use
that despite its deficiencies. Thus we have Windows.

 

[Mike Meyer]

I don't think that's correct. Having nothing can be better than
having something, because then there's an opening to fill. The price
isn't even an issue: it happens even with free programs. If someone
needs an XYZ program and none exists, maybe they'll write a good one
and release it. But if a bad XYZ program already exists, they'll use
that despite its deficiencies. Thus we have Windows.

They'll only use a commmercial program if they can afford it (or are
willing to steal it). And if it's bad enough - in their estimation -
they won't usee it despite it's deficiencies, they'll act as if it
doesn't exist - and maybe write a good one and release it. Thus we
have Linux and the *BSDs.

<mike

 

[Ed Jensen]

Well, there was some concern voiced at EuroPython that a certain large
software-patent-lobbying organisation wouldn't release the shiny port
of Python that they'd done for their mobile telephone products. Now,
one can either emulate that well-practised foot-stamping routine of
yours...

Has this seriously harmed the Python community? Or CPython? Has it
caused evolution of Python/CPython to stall? Did it have the
unfortunate consequence of causing any CPython code to become closed
source or proprietary?

Show me the harm done.

In another recent licensing spat, some people are apparently unhappy
with one Python-related project's use of the GPL, since the code they
originally contributed to an older, related project ends up being
redistributed under the GPL in the former project whereas the latter
project cannot redistribute the former project's original code without
putting a GPL licence on the distributed work. Now, if the latter
project, with its advantage of having come into existence first had
chosen a GPL-incompatible licence, it's quite possible that they would
have avoided the situation that some seem to bemoan, but then one has
to consider the likelihood that people actually do want GPL
compatibility in their favourite open source projects.

I agree that mixing and matching licenses can be a problem, which is
yet another reason I lament the proliferation of source code licenses.

It's like a whole generation of software developers have become
armchair lawyers, which I find unfortunate. Think how much farther
along free software could be if all this energy and concern weren't
expended on source code licenses.

My point about the freeloading was that business understandably likes
to do it. I don't feel any sympathy for participants in various Apache
projects that are hugely popular in business, for example, if those
participants dislike the lack of contributions from those companies
using their software to make money, because those who founded those
projects made a conscious licensing decision and that decision defines
the kind of sharing (or otherwise) around such projects.

I don't feel sorry for them either, and I don't think they feel sorry
for themselves. And the success of projects like Apache are even more
proof, in my opinion, that heavy handed licenses like the GPL aren't
necessary for the success of free software projects.

So if you're not personally affected, as you claim, why does it bother
you?

Because I think a lot of well meaning software developers writing free
software don't performance due diligence to determine the true
motivation behind, and the chilling effect of, the GPL.

 

[Paul Boddie]

Ed Jensen wrote:

[On proprietary ports of Python...]

Show me the harm done.

We'll have to wait and see what happens. There's a risk that versions
of Python with different semantics or characteristics to the original
could cause the development of parallel communities, instead of
everyone working on/with the same project. The "harm done" is
adequately described by paraphrasing your comment on licences: think
how much farther along free software could be if all this energy and
concern weren't expended on separate and sometimes proprietary code
bases.

Because I think a lot of well meaning software developers writing free
software don't performance due diligence to determine the true
motivation behind, and the chilling effect of, the GPL.

Well, despite your protestations, I think the GPL and LGPL are fairly
easy and safe choices for a lot of developers who know enough about
Free Software (ie. haven't just seen the name and thought "that's the
thing for me"), know what the characteristics of those licences are,
and who don't have the time or legal experience to "performance due
diligence".

Meanwhile, all this "hippie" and "chilling effect" talk is, I imagine,
like having a discussion on software licensing with some cold war
propagandist.

Paul

 

[Ed Jensen]

We'll have to wait and see what happens. There's a risk that versions
of Python with different semantics or characteristics to the original
could cause the development of parallel communities, instead of
everyone working on/with the same project. The "harm done" is
adequately described by paraphrasing your comment on licences: think
how much farther along free software could be if all this energy and
concern weren't expended on separate and sometimes proprietary code
bases.

I think free software/open source has existed long enough and with
enough varied licenses (GPL, LGPL, modified LGPL (see wxWidgets), BSD,
X11, MIT, Apache, etc.) that we'd basically know without question if
less restritive licenses (like BSD) were causing projects to fail vs.
projects that use very heavy handed licenses (like GPL). Apache and
Python are two of my favorite examples, followed by the *BSD operating
systems.

Well, despite your protestations, I think the GPL and LGPL are fairly
easy and safe choices for a lot of developers who know enough about
Free Software (ie. haven't just seen the name and thought "that's the
thing for me"), know what the characteristics of those licences are,
and who don't have the time or legal experience to "performance due
diligence".

To be honest, I don't dislike the LGPL that much. The static vs.
dynamic linking issues bother me somewhat (which is why I like the
modified LGPL used by wxWidgets), but all in all, I can live (albeit
uncomfortably) with LGPL. It seems much more sane. Whereas including
one line of GPL code into your 10,000,000,000 line project can have
disasterous consequences (which I find ridiculous), at least with LGPL
you're only asked to share the changes you've made to that particular
library.

Meanwhile, all this "hippie" and "chilling effect" talk is, I imagine,
like having a discussion on software licensing with some cold war
propagandist.

Sorry for my initial post on this subject being flamey. I must've
been cranky that day, and I'm glad we were able to continue the
discussion.

 

[Christophe]

Ed Jensen a écrit :

To be honest, I don't dislike the LGPL that much. The static vs.
dynamic linking issues bother me somewhat (which is why I like the
modified LGPL used by wxWidgets), but all in all, I can live (albeit
uncomfortably) with LGPL. It seems much more sane. Whereas including
one line of GPL code into your 10,000,000,000 line project can have
disasterous consequences (which I find ridiculous), at least with LGPL
you're only asked to share the changes you've made to that particular
library.

If you don't like the GPL, then by all means, *do not use GPL code !*

Please, I mean, when you use without authorisation some code in your
project, you are in trouble, no matter what licence the code was using.

 

[Mike Meyer]

Ed Jensen wrote:
[On proprietary ports of Python...]

Show me the harm done.

We'll have to wait and see what happens. There's a risk that versions
of Python with different semantics or characteristics to the original
could cause the development of parallel communities, instead of
everyone working on/with the same project.

How are the proprietary forks any worse/more dangerous than the
open implementations of Python when it comes to such things? In other
words, what does the GPL do that prevents this forking?

<mike

 

[Paul Rubin]

I think free software/open source has existed long enough and with
enough varied licenses (GPL, LGPL, modified LGPL (see wxWidgets), BSD,
X11, MIT, Apache, etc.) that we'd basically know without question if
less restritive licenses (like BSD) were causing projects to fail vs.
projects that use very heavy handed licenses (like GPL). Apache and
Python are two of my favorite examples, followed by the *BSD operating
systems.

Python and *BSD are getting far less volunteer development love than,
say, GCC or Linux, and the licensing is at least part of the reason.
Also, numerous GCC ports done by hardware companies (for their CPU's)
have been released under the GPL that would definitely have been
proprietary if it had been permitted. That is not speculation, it is
known from discussions with those hardware companies at the time. G++
(the original C++ front end for GCC) also would have been proprietary.

 

[Steven D'Aprano]

if I owned a company
making profit on software sales (sale =! support) you sign a death wish
for using GPL

Apart from Microsoft, and possibly Quark (makers of Quark Express desktop
packaging software), and perhaps a few console game developers, is there
any company making a profit on software sales?

 

[Steven D'Aprano]

I believe in GPL'ed software - I use it regularly. On the other hand,
I don't believe that it represents the best license to release
software if the goal is to improve the lot of humanity. The
restrictions are on "distribution", not on use,

Why would you want to restrict use?

Perhaps if you wrote an evil program that does evil things, and you wanted
to restrict who can use it. But the best way of dealing with that would be
just to not write the evil program in the first place, which has the
happy side-effect of saving you a lot of time too.

The GPL doesn't restrict distribution. I don't understand where people get
this bizarre view of the GPL from. The GPL *encourages* distribution, by
allowing cost-free redistribution. The only restriction the GPL has is
that it prevents the re-distributor from taking away rights which
were granted to them from the people they redistribute to.

If you don't like that clause, you have two very simple options: don't
redistribute the GPLed software. Or use some other software provided under
a different licence. There is no shortage of developers out there willing
to create new software that can be distributed under whatever licence you
like.

so it doesn't really
keep people from using said software commercially.

Why would you want to stop people using your software commercially? That
seems like a good way of making sure your software languishes in
obscurity. If you did, then obviously the GPL is not the licence you
should be using.

For instance, one
or more of your examples may have been worth developing for internal
use. They then decided there was a profit to be made in distributing
it commercially, and proceeded to do so because they could.

I don't quite follow you. Are you saying this is a bad thing or a good
thing? Regardless, the GPL allows the commercial redistribution of
software. What makes you think it doesn't?

Perhaps you think that "commercial program" is a synonym for "closed,
hidden, secret source code". If so, I suggest you check the dictionary.

Without
the profit motive, they may not have done the extra work involved in
preparing the IP for distribution and doing the distribution.

Personally, I release stuff under a BSD-like license, historically
having included requirements that I be notified of bug fixes, and/or
that I be given copies of commercial software that included my code.

That would make it NOT a BSD-like licence then.

 

[Steven D'Aprano]

BSD/MIT style license is a
good substitute of no license at all.

But that's not true: "no licence at all" means that nobody has the right
to use or copy or even *see* your work. You can, of course, choose to
show them your work without a licence, but unless you give them a licence
they can't legally do anything with it.

Perhaps you are thinking of the public domain, which does not require a
licence, but that is because it is not owned by anyone -- not even you,
the creator.

If you want to release your work with no restrictions whatsoever, then
just put the work in the public domain. Is attribution really that
important to you -- especially when that attribution may be buried deep in
the source code of software which nobody will ever see?

 

[Mike Meyer]

On Thu, 24 Nov 2005 16:00:29 -0500, Mike Meyer wrote:
The GPL doesn't restrict distribution. I don't understand where
people get this bizarre view of the GPL from.>

It happens because people say things like:

If you don't like that clause, you have two very simple options: don't
redistribute the GPLed software. Or use some other software provided under
a different licence.

<mike

 

[Steven D'Aprano]

Because I think a lot of well meaning software developers writing free
software don't performance due diligence to determine the true
motivation behind, and the chilling effect of, the GPL.

It took me seconds, seconds I say, to open a web browser and google for
"gpl" and discover www.gnu.org/copyleft/gpl.html

And such chilling effects they are too! Why, if I use GPLed software, I'm
forced to, er, um, well actually I'm not forced to do anything if I merely
use GPLed software.

I'm not forced to pay a licence fee. I'm not forced to maintain licences
at great cost to myself. I'm not forced to get their permission before
publishing benchmarks. I'm not forced to open up the rest of my source
code to others. I'm not forced to redistribute the program to others. I'm
not forced to contribute source code back to the developers. I'm not
forced to allow the BSA to audit my software if they ask. I'm not even
forced to send the developers a post card telling them how much I love
their work.

And if I *choose* of my own free will to redistribute that GPLed work, or
a derivative work of such, the only restriction is that I may not take
away rights granted to me from those I redistribute to. I'm not even
forced to give the software away for free -- I am free to charge as much
or as little as I wish, so long as I don't charge extra for the source
code (excepting reasonable distribution costs of shipping extra media).

Such chilling effects. That explains why Linux and other GPLed software
has languished in obscurity over the last decade, while companies like
IBM, Novell and Red Hat have flocked to support the much older
BSD-licenced code.

Yes, no wonder you hate the GPL, with all those chilling effects.

 

[Ed Jensen]

If you don't like the GPL, then by all means, *do not use GPL code !*

Please, I mean, when you use without authorisation some code in your
project, you are in trouble, no matter what licence the code was using.

I'm not sure why you felt compelled to state the obvious...

 

[Ed Jensen]

Python and *BSD are getting far less volunteer development love than,
say, GCC or Linux, and the licensing is at least part of the reason.

I disagree. I believe *BSD gets less volunteer development because of
some legal wrangling in the early 90s that didn't affect Linux.

I believe GCC gets more volunteer development than Python because C
and C++ were (and are) much more widely used.

Also, numerous GCC ports done by hardware companies (for their CPU's)
have been released under the GPL that would definitely have been
proprietary if it had been permitted. That is not speculation, it is
known from discussions with those hardware companies at the time.

Even if this is true, GCC would have continued to exist. Just because
an entity takes some open source code and places it in a closed source
product, the original open source code continues to exist.

Frankly, I suspect those hardware companies would have relented their
decision once they realized it was harder to keep re-integrating their
code into newer GCC releases, than it was to just release the code.

G++
(the original C++ front end for GCC) also would have been proprietary.

I'm not saying you're wrong, but since you're providing no evidence,
I'll remain skeptical about this claim.

 

[Steven D'Aprano]

Whereas including
one line of GPL code into your 10,000,000,000 line project can have
disasterous consequences (which I find ridiculous)

If you think that's disastrous, just try using one line of proprietary
code in your 10,000,000,000 line project without permission.

Or for that matter, one line of BSD code without living up to your
obligations under the BSD licence.

That's what is is really about, not the presence or absence of lines of
code. If you can't or won't live up to your obligations under the licence,
then you have no business using the code in your project.

And no, it isn't too obvious to mention -- all the nonsense talk about
the GPL being viral misses the point that *any* unauthorized code can and
will poison your entire project if you get caught. Anyone who thinks the
GPL is unique in that regard is deluding themselves.

 

[Alex Martelli]

Why would you want to restrict use?

I can think of many reasons. For example, if the author of some piece
of software strongly dislikes [category A], they may wish to restrict
firms or people in category A from using their software, because said
authors believe the use would be helpful to A's purposes and thus
damaging to humanity. You can easily imagine various different authors
for whom A could be "companies which build weapons", "companies which
belong to the RIAA", "abortion clinics", "schools which teach evolution
and not creationism" [or vice versa], "walmart", "the American Nazy
Party", "the American Communist Party", and so on, and so forth.

Licenses excluding uses by some specific category may be legal
(depending on jurisdiction and exact definition of category) but they're
definitely not open-source, by definition of the latter.


Alex