GPL 3.0 to Penalize Google, Amazon?

Michael Ferris writes "Is this the start of a shakedown by the GNU folks? Michael Singer writes that Eben Moglen and the folks rewriting the GPL are looking at a proposal where companies would be required to pay money if they use GPLed software, even if they don't redistribute the software." From the article: "The current version of the GPL, which was last updated in 1991, fails to trigger the open source license if a company alters the code, but does not distribute its software through a CD or floppy disk...the [current] rule does not apply to companies that distribute software as a service, such as Google and eBay, or even dual-license companies like Sleepycat."

Future versions of the GPL

If I am ever to write something worthy of releasing to the world (and not just something I am playing around with), I will explicitly specify which version of the GPL I am releasing it under.
Call me paranoid, but I wouldn't want even the faintest chance that some nasty corporation managed to litigate itself in the position of being able to release a future GPL version, as in bold below:

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

That's handing control of the licencing of your code over to whoever is allowed to write GPLv1851, if I'm reading it correctly.

Re:Future versions of the GPL

That's handing control of the licencing of your code over to whoever is allowed to write GPLv1851, if I'm reading it correctly.

I'm worried about GPLv1984 myself.

Re:Future versions of the GPL

RMS is a madman. Fortunately, he's our madman.
He needs to be watched closely (to prevent blunders like the GFDL), but, he is well-known to have good intentions.

The problem is, if something bad happens to him, it's possible that whatever members of FSF will have the deciding say will push the GPL in a completely different direction. Whoever controls the FSF, controls the vast majority of GPLed software.

I'm not paranoid enough to label FSFians as possible traitors who would follow whoever shakes the purse, hell no -- I have quite a bit of faith in them. However, they may do any modifications to the licenses of software they don't own the copyright to -- it's a huge power. It's dangerous to leave such power in the hands of people not protected by insanity.

Re:Future versions of the GPL

Well, I do have software distributed under the GPL, so I want to talk specifics...

Here's the full section 9, a portion of which you quoted:

9. The Free Software Foundation may publish revised and/or new versions of the General
Public License from time to time. Such new versions will be similar in spirit to the
present version, but may differ in detail to address new problems or concerns.

Each version is given a distinguishing version number. If the Program specifies a
version number of this License which applies to it and "any later version", you have
the option of following the terms and conditions either of that version or of any
later version published by the Free Software Foundation. If the Program does not
specify a version number of this License, you may choose any version ever published by
the Free Software Foundation.

So my question becomes this: What determines whether or not you specified a specific version of the GPL? Most of my comments and the readme file say "Licensed under the GNU General Public License" (no version mentioned), but then included with the distribution is a copy of version 2 of the license. Does that imply strongly enough that version 2 is the specific license under which the software is distributed?

Or do I need to go make some changes and do a commit....

Re:Future versions of the GPL

A few points for sanity here:

1. No one can force a company to abide by the rules of a new version of the GPL for software the currently have under an old version. They can CHOOSE to apply the new version IF the author used the standard boilerplate license notice, but they can also CHOOSE not to
2. The article specifically states that there is no GPLv3 and they're not officially comenting on what they are considering for it when it does come out
3. The guy commenting is saying what he would like to see so that he can drop his unique license. That's fine, but it's not official word
4. If this were to happen in the doomsday sense, everyone woudl immediately fork old copies of the programs that they have the option to apply the GPLv2 to, and continue to maintain and license them as such.

They deliver HTML.

the [current] rule does not apply to companies that distribute software as a service, such as Google and eBay

Google and eBay distribute HTML. That HTML is created by software that uses GPL code. So if I modify a GPL Office Suite, would I have to distribute the code if I email someone a document I made with it? Seems like a bad idea, in general.

I guess people could fork the GPL2.0'd code if the software developers switched to GPL 3.0

Re:They deliver HTML.

Google and eBay distribute HTML

Hmm. I could swear that the Google appliance in my rack, and the Google toolbar on my desktop weren't just hunks of HTML.

Re:They deliver HTML.

Has it every been proven that the google appliance is a Linux box, because Ive seen one in the flesh and played with it on a network, and it most certainly looks like a unix box of some description (nmap identified it as a FreeBSD 4 server among other things) as of 6 months ago.

Just because they use Linux in the Googleplex doesnt mean they use it everywhere.

The GPL must change or die.

Google and eBay allow end users to run programs on Google and eBay servers that create HTML for the end user. If I go to Google and do a search, INPUT occurs on my computer, PROCESSING occurs on Google's computer, and OUTPUT occurs on my computer.

That's not the same as you distributing a document you created. In that case, INPUT occurs on your computer, PROCESSING occurs on your computer, and OUTPUT occurs on your computer.

The question is: What's important in determining if a program has been "distributed"? Clearly if I give you a CD with the program and you run it on your computer, I've distributed the program. Clearly if I take GPL software and modify it and I run it on my computer only, I have not distributed the program.

But if I take GPL'd software, modify it, and then let YOU run it on my computer...

Did I distribute it, or not?

Your immediate answer is probably "Who cares?" But now what if I charge you to run this modified program on my computer, for example, by charging a fee if you use my auction program? Now I'm using software provided by the open source community for my financial gain, but not returning the modifications I made to that software to the community.

The extreme of this problem is that eventually, the internet becomes so fast and clients become so dumb that software is never "distributed" at all. I take an open source office suite and then modify it. According to the GPL, if I then sell that software on CD, or by download, so that people can actually run it on their computers, I must provide the source to it as well. But what if I'd rather just make money off of the GPL'd software I've taken without giving anything back to the community?

Well, then I just put the software on my own server, and instead of selling CDs or downloads, I let people provide input and receive output over a remote connection to the program running on my server.

And wala! People can modify and essentially provide GPL'd software without having to provide source.

This doesn't sound like a good idea

I think requiring companies to pay is a big time mistake, and to me, it
hails back to the days of the old BSD style licenses. The GPL works
so well now, precisely because it is unobtrusive. My company runs
GPL'd software because we are able to use it and make modification
without either redistributing the source code and we aren't required
to pay for that ability.

Requiring people to open all their changes or pay for them will put a
lot of businesses off when it comes to dealing with GPL'd software. I
don't think that is a good trade off to make, and I don't think it
will be healthy for the open source community in general.

A move like this will make the newer BSD style licenses and / or
licenses like the Python license much more attractive imo. Now that
open source is finally turning the corner, and solid technologies are
finally moving into the enterprise, why would we even entertain making
changes that will certainly hamper open source adoption?

This isn't a consistent position in my opinion. If you are
developing free software, it should remain free. The idea that you
can get someone hooked on software, and then pull the rug out from
under them and start charging them is ludicrous. If this were to
happen, I can honestly see a major fork in the GPL happening.

Re:This doesn't sound like a good idea

I'll go farther: This is an incredibly stupid idea. It will only convince businesses that the Linux/free software folks really are the communist, business-hating zealots that they're painted as.

If you're not distributing your work, there's no reason why you should be forced to open your code or pay some silly fee.

I hope this whole story is a troll...I really do.

Re:This doesn't sound like a good idea

"The idea that you can get someone hooked on software, and then pull the rug out from under them and start charging them is ludicrous."

These implications of the proposed GPL3 are certainly troubling. How is this different from "evil commercial vendor lock-in"?

I'm not trolling...just hoping that this interpretation of GPL 3 is wrong.

Let's get more detailed

But who gets paid the money and who determines how much? Can rates increase as GPL'd software develops? Is money paid to Richard Stahlman [sic] to distribute through the FSF? Seems an easy way to get around this anyway would just be to have small closed-source scripts that would only be called from the GPL'd code. After all, what's wrong with that? (so long as they're not redistributing their code, just letting use of it as a service)

Im speechless.....

"If you look at the market, Yahoo, eBay, IBM, Amazon, Google have all sunk millions into the GPL infrastructure," Olson said. "Not only are we changing the rules, we are changing them retroactively. With the new way, it lets the customer pay with either their source code or with their wallet."

Basically, in any other language: 'Now youve had time to build a good infrastructure on the current rules, prepare to be shafted'. If this comes to pass, then in my mind they are no better than Microsoft changing EULA terms with a service pack. Now that there are major companies with an infrastructure built on GPL software, the FSF are looking to essentially move the goalposts and if this is applied retroactively to current code (which from the articles wording I think it will be) then I personally think that its going to do more harm to the GPL community than benefit it in code donations as companies scramble to move away.

Please someone tell me that they cant do this retroactively, that its impossible under the current GPLv2 terms.

Re:Im speechless.....

Please someone tell me that they cant do this retroactively, that its impossible under the current GPLv2 terms.

Your wish is my command. If you look at source code to any GPL-licensed program, you will see something like:

// This program is free software; you can redistribute it and/or
// modify it under the terms of the GNU General Public License
// as published by the Free Software Foundation; either version 2
// of the License, or (at your option) any later version.

Hence any code that was licensed under the GPL before version 3 is released may still be used under the terms of the GPL version 2, at the option of the recipient, not the author. Actually, a number of current projects, including Linux, are licensed GPL-2 only and may be impossible ever to convert to a higher version (permission would be required from too many people to reasonably track down). Hope that helps.

Horrible implications.

Just the implication of this terribly and inaccurately worded headline will be devastating to open source. This article is intentional flamebait. Zonk, please try to be a bit more responsible with the articles you post.

Regardless of what GPL 3.0 turns out to be, developers are not forced to use it. They can continue to use GPL 2 if they wish, just as they can choose to use a BSD license, Apache license, creative commons, or any other license of their choosing. Furthermore, software that has already been released under GPL 2 cannot be retracted, it remains available under GPL 2 forever.

Submitter mischaracterises the change.

The submitter isn't clear about the fact that this would not apply to everyone who changes software for commercial use but does not redistribute. This applies ONLY when "redistribution" of the software sort-of-occurs, because the software is used to provide a service. For example: any open source software Google uses in its search engine interface does not count as "redistributed", even though it _interacts_ with users of Google, because no actualy _binaries_ were shared with those users. For another example, if I modified the GIMP and then let people use my modified version over the Internet, but not on their machines, I would not have "redistributed" my modifications. This is considered by many to be a "loophole" to be closed.

Re:Submitter mischaracterises the change.

Right, so for example, a friend of mine works for a financial firm which he reports makes use of (and even contributes to) OSS projects. That firm provides a service based on these products (both electronic and off-line transactions that they perform as part of their core business). If they attempt to make this retroactive, I assure you that the world will come crashing down on the FSF. Thousands of firms around the country will sue them OVER NIGHT, and honestly, I'd be more likely to donate to the defense of those firms than the FSF (regardless of the fact that I'm a huge fan of the GPLv2).

This is deeply irresponsible. Any project that ships software under the GPL is going to be spinning their wheels for months over this, and the Microsofts of the world just got a huge weapon to use against OSS usage. After all, now they can say that using GPL software not only costs you in terms of the usual TCO metrics, but there's a potentially hidden and as-yet-unknown cost that can be applied to retroactively!

Grrr!

Where did this mindset come from?

Why is everyone always saying "Software should be free, unless you're a business, in which case, get your checkbook"? The GPL shouldn't mandate anything except that code be made available. I think corporations should be entitled to the same rights and privileges as private citizens.

Re:Where did this mindset come from?

> I think corporations should be entitled to the same rights and privileges as private citizens.

Like voting? And do corporations get the same negative sides as private citizens, like going to jail? If you won't send a company to jail, and give it a vote, you can't equate them to private citizens

Call it FUD

Frankly, I don't think this is even the case. Yes, the FSF has been talking about making a GPL 3.0 for a while now, but the proposal they're offering makes no sense.

The problem is that it infringes on one of the four freedoms, specifically the freedom to use. If such a provision were to find its way into any license that made it so that companies and individuals were not subject to the same terms, the license would be both discriminatory and non-free.

Simply put, this is somebody making FUD about the GPL. Don't buy into it for a second.

It's not in print yet... it doesn't exist

I heard the next version of the GPL will require everyone to throw pies at Bill Gates.

At least, that's what I heard anyway.

When whatever changes come up, they will be reviewed and we can rant and rave about it at that time.

Nothing is founded, no concrete written agendas were tossed out by the article and all we have is a hypothetical situation that would be very different from the current model.

Like many have said and will say, it's usually a good idea to specify the version of the GPL in which you release software. Unless you really don't care what kind of changes are made at a later date.

I'm not saying all of this isn't going to come true, but at this juncture we could very well be required to throw pies to comply with the next redistribution agreement.

Software for the Rest of Us

That new constraint would be a totally different principle for the license. The current principle is that if someone gets an executable from you, they also can get the source code, just as you got the source code from which you made your executable. Just using the source code, or customizing it for your own use, doesn't require distributing the source. The new principle would be requiring anyone who customized the source to release all customizations.

This will stop many programmers, many of the best programmers, from using source code under that license. We could no longer keep any of the value of the software we created to ourselves. What is a "web service"? Is my email-processing CGI a "web service"? Any software in the same workflow as any other software under this license would have to be released. So many developers won't make small customizations, because that would force us into the source code distribution business, with all its overhead. Or we might just ignore that provision, or the whole license, en masse.

The GPL is successful because it is a fair contract, even though it's revolutionary. Its enforcement teeth are rarely tested, because it's so close to an equitable agreement among peers. Which has resulted in lots of value contributed by profit-driven organizations, despite the claims of many that the license is anticapitalist. Upping the ante, to require private customizations to be published, could stop the rising tide of acceptance that is pushing GPL to be the default, and any proprietary license to be radical. And then the caution it would inspire: investing in GPL'd software might force acceptance of ever-more demanding licenses, like a GPL4.0 that required redistribution of even software that wasn't changed at all, just to get users "to pull their weight".

The GPL2.0 isn't broken. Let's not "fix it" in a way that could destroy its success, and our chances to benefit from one another's work without onerous burdens.

Re:Who what when where?

Why you'd pay it to the FSF, of course. They'd administer the money, funding projects as they see fit. Kind of a Politburo for the Software Community.

BTW, when does Stallman's Macarthur Foundation Grant expire?

For the humor impaired: It's a joke son.

Re:Who what when where?

The payment model could be based on a simple royalty each time each subroutine is accessed. The easiest way to implement this would be an opaque wrapper on all GLPed APIs, which in turn could be easily implmented in a new version of gcc.

The wrapper code could then count the number of calls, the function name called, and the company calling them, and this in turn could append a text file that is stored somewhere and automatically sent once a month to the FSF which will then use the file to send the company a bill. The FSF would then distrubute the money to each of the authors of the given API calls.

I will volunteer to make the necessary modifications to gcc. This of course will require all GPL'd code to call my APIs, which will of course be the first to implement this new system.