That says more about you than it does about the quality of the analogy.
Casablanca is still a great movie, you can watch it "in 2005"; but it's not a movie "of 2005".
But the iPod *is* a product (specifically, a brand with current products) of 2005.
The rules for the Academy Awards are not the same as the rules for PC World's list.
and like movies, sequels rarely win awards.
That's one of the dumbest things I've read all day. Look at the list and notice all the products which are "sequels".
In fact, the iPod that's on the list is a "sequel", and clearly it won.
the point was it's not the original 2001 one,
No one said the original iPod should or even could win. You know why? The original iPod hasn't been an contemporary product for years now.
The dateline of the iPod Photo article: Friday, November 19, 2004.
Your premise, that a product that was not released in 2005 is ineligible for a "Best Products of 2005" award is moronic, and the fact that products that debuted before 2005 (Google is on the list, you idiot) doesn't seem to clue you in.
You're defending a lost position and you know it. Your reasoning is not self-consistent, nor is it even consistent with the article.
Don't bother replying, you'll just make yourself look more foolish. The iPod (in various flavors) *is* a product of 2005, it *is* on the list, and it is *not* a movie. These facts seem fairly basic, and quite easy to grasp.
You can still watch Casablanca at your local revival cinema, so why doesn't it get an Oscar every year?
Why would you think the rules for winning an Academy Award apply to a PC World "Top 100 Best New Products of 2005" list"?
I mean, seriously, are you trying to claim that since Casablanca can't win an Oscar for 2005, that the iPod can't be considered a "Best Product of 2005"? One's got nothing to do with the other.
That's the "new, 2005" video version, I think.
There is (currently) no '"new, 2005" video version'.
You seem to be under the mistaken impression that this is a "Top 100 Best New Products of 2005" list (which it isn't), and not a "Top 100 Best Products of 2005" list (which is what it actually is). Last I checked, the iPod is not only still a product in 2005, but is a very popular and highly acclaimed product.
There *are* better things out there than the iPod. How is this surprising?
It's surprising given that the iPod usually takes a couple of slots in the top ten on lists like these.
And when you have such a generalized list...well, you will always get strange results. What was the criteria for determining a product that would make the list?
Yeah, that's just it. Why would the iPod fare lower than a hard drive? Or a Rio mp3 player?
I mean, if I were going to bet on whether an iPod or a Rio would rate higher on a "best products" list, I'd bet on the iPod. This isn't a earth-shattering surprise or anything, just a "huh, didn't expect that" sort of surprise.
OK. If it's so obvious that a person is 'using' the backend program, then the program is clearly being distributed to them (otherwise how are they using it?), so I struggle to see how the current GPL fails to cover it.
Distribution is another one of those words which has an objective meaning. Distribution means the person is receiving the binary code, for example on a physical disk, or via download. The web surfer doesn't receive the program, they just operate it.
So, no, the program is not "clearly being distributed to them".
This is similar to how you can use a TV from across the room, without actually touching it. Remotely operating something has been around for quite some time now.
I did "see" that, and as I've made quite clear, my view is that you, as website visitor, are not the 'user' of the backend server but the 'user' of the generated pages.
...
You, on the other hand, seem to be wilfully ignoring my opinions,
...
I simply don't believe that the user of a website is any more the user of the backend application that generates it than the reader of a book is the user of the word processor that generated it
The definition of a "user" is not subject to your views, opinions, or beliefs. It's a well-defined, and objective term. The GPL is not contigent upon your beliefs.
It's so simple as to defy belief you can't understand it. *You*, the web surfer, are the one who entered data into the web form and clicked the "Submit" button. That's the interface by which you operate the remote program.
This is analogous typing a command in a terminal window and pressing "enter".
"But," you cry, "the web surfer isn't the one sending the command, it's Firefox or IE!" Are you so simple you are confused by a layer of abstraction? If you load a tiff in the GIMP, the GIMP calls libtiff to process the image. It does this by your request. This makes you a user of libtiff as well, even though you never personally, nor directly, called that library.
With a web browser, you are personally instructing and interacting with a remote program. Which is to say, you are using it. All your views, beliefs and opinions cannot change that fact.
(or users of an application which was compiled using a modified, GPL-based compiler, and if I understand the GPL correctly, in this instance there is currently no compulsion to release the modified compiler).
Right, and the new GPL won't require that either, because running a program that someone else compiled doesn't make you the user of the compiler.
From TFA "In a broader sense, Arnold believes Google is building a "patent fence around search" technology as the firm moves to codify its unique competitive advantage."
Is this good or bad?
Guns are bad.
Given that guns exist, if the police use guns to fight something worse than guns, it's (potentially) good.
Patens are bad.
Given that patents exist, if a company uses patents for a greater good, it's (potentially) good.
While there is no doubt a contigency in the slashdot crowd that has an irrational need to be absolutely consistent in a superficial sense, there is a valid and consistent logic to being opposed to the existence of patents while, accepting the fact that they do exist, supporting their use for good causes.
I personally don't see anything in the *original* intent of the GPL (not just in the subtleties of the wording) that is being breached by using GPL
The preamble of the GPL begins with, "The licenses for most software are designed to take away your freedom to share and change it. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users." and continues on from there to further clarify the intent behind the GPL.
If you didn't "see" that in the GPL, you clearly haven't even read the GPL.
If Stallman wanted to extend the purpose of the GPL
He doesn't want to extend the purpose of the GPL. As you can see above, he wants to protect the freedoms of "all its [free software's] users". "All" includes those who use software remotely.
My view is that when browsing a website, you are 'using' the HTML and any javascript thats running.
Nice try, but it's not the 'using' of the HTML output that is in question. It's the issuing of commands (such as via HTTP), which makes you the user of the remote software. The data sent via HTTP (such as HTML, JavaScript, images, and so on) is the output of the software on the remote server that you, the web surfer, are using.
No. I'm flabbergasted that anyone can see this as tightening a loophole as opposed to a significant change to the scope of the GPL. As I think I've already mentioned Stallman is free to do what he wants in creating a new GPL, but trying to claim it's simply tightening a loophole is just plain wrong.
I've never said it wasn't changing the scope of the GPL. In fact, I've stated that that's exactly what it is *many times*.
You can't read English, can you? The loophole isn't between the GPL and the usage in question. It's between the intention behind the creation of the GPL and the use in question.
Quit being so dense.
Read carefully (I know you won't, but I type this anyway):
1. Stallman wishes to promote four freedoms. 2. He creates the GPL with that goal in mind. 3. The GPL does this splendidly. 4. A relatively obscure, yet absolutely valid, usage of the GPL becomes common which allows people to violate the four freedoms.
That's the loophole. It's between 2 and 4, not between 3 and 4. I've pointed this out about a million different ways. If you can't see that, and you can't stop parroting your original, "but it doesn't violate the wording of the GPL", then my participation in this thread is over.
The whole reason loopholes exist is because people are able to find ways to violate the intent of a thing (a law, for example), while complying with it to the letter.
What is the 'browsor' (my term for a browser operator) using, other than his browser?
It's really simple. The user is the one who uses the program. For interactive programs, the user is whoever is clicking the buttons, typing the keys, etc.
If you are using a program which in turn sends commands to another program, *you* are using the second program as well.
For example, if you were to put a program online which creates images of the current sky (astronomy/planetarium software), and the user is presented with a form to enter their coordinates, and a few other parameters, the web user is, in fact, using their web browser to control the astronomy program. In other words, they are using the astronomy program.
The same applies with Google. If Google is using modified GPL'd code to dynamically serve up maps, the user who directs their browser is using that GPL'd code.
The person in question is not the 'user' who is running Apache, perl, Python, Ruby, or whatever on the server. That person is 'anonymous', 'guest', or 'nobody' on the server.
That is a user id. A user id is not a person, it's a number. It's nonsense to say "anonymous" or "guest" or even "billg" is the user. The user is the actual person using the program (and who may or may not be using some given user id or another). If you are logged into your X11 session as "themonster", and you have the GIMP running, if you hand me the mouse and keyboard to do some photo editing, I am the user of the GIMP, not you.
Now, of course, licensors are free to ignore this distinction, and create a per cpu license, or per instance license, or whatever. But the GPL is a license intended to provide a certain set of freedoms to the user, so it's logical to apply this to users of remote software as well.
For the millionth time, I fully agree that there's no loophole between the wording of the GPL and the use that's under question. You can quit saying things like, "What I'm saying is that I see nothing in the original GPL that suggests there was ever any intent to cover this sort of scenario."
Read carefully: The intent of the GPL is to protect four very specific freedoms. The GPL, as worded, fails to do so. It's a way for people to comply with the letter of the contract while violating its intent--which is to protect the four freedoms. Please don't respond with, "but the use in question is not covered by the GPL," because I've already made it abundantly clear that I agree it isn't.
The point I'm trying to get across is that 'web-enabling' a system doesn't seem to me to be anything particularly special (and often doesn't involve doing anything specifically to the app at all). People have been interacting with companies' back-end systems through a variety of direct and indirect means for years, and none of these have been seen as 'distributing' the app (for the purposes of the GPL), and I don't see why a website is any different.
That's because you are looking at the issue in the wrong way. RMS isn't primarily concerned about distribution, he's primarily concerned about freedom. The distribution aspect of the GPL is the means by which he's attempting to legally protect the four freedoms he believes should be protected.
When you look at it simply from the point of view of distribution, you are correct that the GPL doesn't really apply to web services, and you know what? RMS agrees with you. That's why he's working to update the GPL. Absolutely everyone is saying that the GPL doesn't apply to web services. Google is saying it, RMS is saying it. I'm saying it. Yet still, you jump up and down screaming, "but it doesn't apply to web services!!!" as though you're arguing with someone.
Basically, I'm struggling to understand how (or why) the licence requirements for an application can change the second the app becomes involved in a web transaction,
That's up to the license. Most licenses don't make that distinction, such as the current GPL. Other licenses do, such as the license for IIS, Exchange, SQL Server, etc.
To me this feels a fundamentally different act to that of releasing an application binary for someone else to pick up and install.
That's because it is a fundamentally different act. No one is claiming otherwise. The commonality is that the four freedoms in question can apply to both situations.
You're right that people are free to choose whether they will use GPL3 code in the future, but it is a big deal. I know that the next enterprise app I design will not include any GPL3 code if it means that as soon as I web-enable it, I'm forced to publish my private business logic(whereas I've got no problem with the current GPL as I've got no intention of making my apps available to third parties to install).
Assuming you do hardcode your business logic, and assuming you do wish to keep it secret, then you are fully free to do so. The purpose of the GPL isn't to make free software palatable to someone who wishes to keep their software secret. The GPL is designed with the primary intention to allow for people to create software which provides the user with four specific freedoms.
You're confusing the intention of the Open Source community with the intention of the Free Software Foundation. They have two very different goals. To the Open Source community, it *is* a big deal if their license makes it hard for you to use their software. To the FSF, the big deal is when you are allowed to supress the four freedoms they wish to promote.
It all depends on what you want. You sound like your are more of an Open Source kinda guy, so it's no big deal that you don't like the GPL 3, since you aren't the target audience.
You seem flabbergasted that the FSF would want to update the GPL to cover web services, but if you understand their intention, it's quite reasonable.
Re:Integration versus Bundling, Choice
on
Pepping Up Windows
·
· Score: 1
So where am I getting at? The key reason why Microsoft got in trouble was it's INTEGRATION (IE as a part of the OS) and LACK OF CHOICE (media player installed by default) and not the fact that it was bundled on the CD.
The problem wasn't the integration so much as the fact that MS, by virtue of providing the OS used by over 90% of all PCs, was able to (and in fact did) use that product to force other products on the user. The problem *was* the bundling.
In retrospect, the bundling issue sounds incredibly lame. What modern OS doesn't include a media player or a web browser as part of the/a default install?
It's silly. OS X bundles Safari and QuickTime. GNOME bundles Totem and Galeon. KDE has Konqueror and whatever their media player is. Virtually all distributions bundle Firefox and xmms. Even BeOS and Amiga OS have bundled browsers and media players.
MS bundles IE, but I used Netscape exclusively until IE 4, and I now use Firefox. MS bundles Media Player, but I still use VLC, QuickTime, Real, iTunes, and about a million other media players.
My problem is, as I mentioned, that as far as I can tell, the GPL talks about the distribution of the software, not the use of it.
That's correct. I've already said that. The loophole is not between the wording of the GPL and the use of the GPL, those two things are in perfect sync, as regards to web services. The loophole exists between the intention behind the GPL, which is not about distributions, but about promoting four very specific freedoms.
I'll state this very clearly:
You are correct that there is no loophole being exploited related to the wording of the GPL.
There is, however, a loophole related to the intent of the GPL.
Quit bringing up the wording. We agree on that aspect, and so does Stallman. If he didn't agree with that, he wouldn't be so set on changing the GPL.
Say I've got a (non-web enabled) account management system, which includes some GPL code, and some proprietary business rules. My understanding is that under GPL I would not have to publish the source code for this app.
Correct. And "proprietary business rules" sounds like your talking about configuration of a compiled binary (similar to rules in a mail app). You've not changed the source code, but for the rest of this example, I'll assume you have.
Now say I build an website on top of it, allowing the users to view and update various aspects of their accounts (such as changing their address). At this point, would I have to publish the source code for the backend account management system?
Under the current GPL, no. Under the new GPL, assuming what we hear of it actually comes to pass, yes. That's exactly what the authors of the GPL intend.
Stallman believes users of software should have four specific freedoms, and he has written a license specifically designed to promote and protect those four freedoms.
I don't know if you posed this scenario because you didn't know the answer, or if you intended to show a scenario which you find undesirable (it can be read both ways). But if it's the latter, I don't see why someone would really be bothered by the fact that they have to provide access to the source code of software they got entirely for free. But as with all licenses, the author of the license (RMS and Moglen, in this case) are free to include any legally allowable terms they desire, and you are equally free to accept or decline those terms.
This is not a big deal, and if you are currently using software licensed under the GPL, you will not have to abide by version 3 of the GPL if you don't want to. In other words, if you are currently exploiting the loophole in question, you will be allowed to do so forever. The new terms only come into play if you accept software under the GPL v3 (or later) license.
That's very different to a "loophole" where the intention of the licence is extremely clear but people are using clever tricks to avoid it.
The intention of the license is very clear--it's to promote four very specific freedoms. In the opinions of RMS and Moglen, the current version of the GPL fails to ensure those freedoms in a specific set of circumstances. This is due to a loophole--a loophole not between the wording of the GPL and this specific set of circumstances, but a loophole between the intention of the GPL and that set of circumstances.
To my knowledge, the current GPL talks about copying and distributing, not using.
Copying is where the GPL applies, because it's based on copyright law. Distribution is where they chose to focus because:
1. This keeps the GPL simple (too complex and unintended consequences are more likely) 2. Distribution was the most prevalent way users became users.
Now, it's increasingly common for a person to become a user (as per the opinions of RMS and Moglen, and their opinion is what applies regarding the intent of the GPL, as it is their document) without having the software they use having been distributed to them. Therefore, in order for the GPL to more accurately address the intentions of its authors, it will need to be updated, which is exactly what they are in the process of doing now.
Indeed it would - so if I used my program to write static HTML which I then published, would this program require me to publish the source?
I'm fairly certain they do not intend this, and that the new GPL will not require this. If you read the pages on the FSF site, it will become clear that they believe users should have the right to use modified GPL'd software without distributing their changes if they so desire. Your static HTML example would fall into this category.
It's only because web surfers are, in effect, using the software that this question has come up.
If you use a program to create a static html page, you are the user. If I click a button on a web page, the program that is run by my click is now being used by me, and thus I am the user.
At least, as far as RMS is concerned, as I understand it.
Yes, by limiting the freedom of people to use free software for purposes Stallman doesn't like.
Absolutely wrong. Stallman doesn't like Communism, but his license doesn't say the Chinese can't use emacs to draw up the death-sentence for some dissident.
The *sole* purpose of the GPL is to promote the FSF's Four Freedoms. Nothing more, nothing less.
It's an irony, but inescapable fact, that to promote freedom, one must actually limit freedom. It sounds illogical, but it's not a matter of logic, it's a matter of reality. For example, if you don't have laws that limit someone's ability to kill someone else, you have less overall freedom because the threat of being killed can be used to force people to do what you want.
Similarly, the GPL limits certain freedoms (more specifically, it outlines certain responsibilities, and limits your ability to add further licensing restrictions to a piece of software) in order to promote more freedom overall. Of course, just as in the real world, it's possible to go overboard and limit freedom in such a way that is not rational. I'm sure you can think of some laws right now that do that. The GPL, on the other hand, does not do that. The GPL is simple, is predicated on four basic freedoms, and is fully consistent with those four freedoms.
Plus, let's not forget what happens when you make free software too difficult for commercial entities to use... that's bad for the free software world, because if it happens too often, people will standardize on something else and free software will become marginalized.
RMS's goal is not for all software to free. His goal is for free software to exist. He has stated many times that you are perfectly free to use non-free software all you want. ESR is the one who's more concerned with promoting a system he sees as superior for practical reasons (read The Cathedral and the Bazaar), and while he's a libertarian, he really isn't primarily focused on promoting freedom, he sees that as a nice side-effect.
So, stating that the new GPL may make some companies not use the GPL is not a counter argument because it doesn't have "companies use GPL software" as its primary goal. If you want that, the BSD license is more in line with that goal. You (and everyone else) is free to chose whatever license they wish for their own fully-owned code.
When he says "working", you and I both understand that he means "working insofar as allowed by necessary security restrictions put in place to prevent abuse of the service." But the GPL is a legal document, and unless that's explicitly stated, it could be interpreted as meaning no restrictions allowed, period. And that's how corporate lawyers and suits will interpret it - encouraged, no doubt, by the various FUD factories.
It's a bit premature to parse words that aren't the actual license. The GPL currently requires one provide access to the source code via a network service (like ftp) or on physical media, and no one has argued that the ftp site has to allow itself to be overburdened, or that a person has to ship out CD's immediately. What's likely to be the case is that the idea of "reasonable" will be applied. In other words, what does the industry consider to be "reasonable" access via ftp? Or a "reasonable" rate of delivery of physical goods?
Maybe it'll be used as FUD, but I really doubt it, as I've never heard (until you brought it up just now) anyone imply the GPL would require you to overtax yourself or your equipment in order to provide access to source code.
I think you're wrong. The user of a program is the one that downloads installs/authorizes code execution on their system.
Your definition of "user" is flawed. A user is (it's in the actual word) the one who uses the program. That's it. One doesn't need to download, install, or authorize a program to be a user, one merely needs to use the program. That often entails downloading and such, but doesn't require it.
The owners of a website use a web server and web apps to deliver output to the persons browsing the site.
No, they don't. A computer delivers the output. The owner of the website provides/allows access to the services, the services provide the output.
Likening use of HTTP to linking a library is wrong, and IMO idiotic.
Good thing no one is doing that.
I find the uproar over this quite amusing. If MS, Apple, IBM, etc, can restrict network use of their software via license, why can't the FSF? The fact is, they certainly can.
There are really only two questions:
1. Is this in line with the ideals of the FSF? I believe it is, but more importantly, RMS thinks it is. 2. Will this cause unacceptable problems? It's up to those who are in a position to accept the license to determine whether it's acceptable or not, which requires an actual license to begin with. As for whether it causes problems is up to RMS's and Moglen's ability to draft a useful license. They have shown proficiency at this in the past, so it's a bit unfair to assume the worst.
Not in the sense you're talking about, of course not. But, if you're legally bound to leave a particular service open and unthrottled, it could become a potential vector for DDoS attacks. There need to be provisions to allow a site admin to place limits on the number of simultaneous requests, the frequency of requests from a single client, and other such limits to help prevent abuse.
That's a very odd thing to say, and I don't quite understand where you go the idea that webmasters or site admins would be required to leave a service open and unthrottled. In fact, that would go counter to the FSF's "four freedoms".
It will be backed by legalese (though a human friendly dialect where possible), since last I understood the nature of a license is legal, and they want to be legally clear and unambigous about their license. Again, feel free to prove otherwise.
Prove what otherwise? That the GPL is a legal document? Don't be silly. That's not what you implied. You implied that it will be "very dangerous" because it "is backed by much legalese". Without having seen it, you aren't in a position to judge how "much" legalese it will require, and how "very dangerous" that will make it. I can't disprove your fears, but it's not my requirement because I'm not making the claim, you are. My claim is merely that you don't have the necessary facts on which to found your statement, which I can prove by the simple fact that the GPL v3 doesn't even *exist* yet.
The problem is that making more licenses for shits and giggles is a bad idea
The GPL v3 isn't being created for "shits and giggles", it's being created to address a very real concern.
and over time will be an extra burden at the very least.
I agree, which is what I meant when I wrote, "that incompatible licenses create friction? Of course they do." I thought that was fairly clear.
You bring up valid concerns, but you can be very certain that RMS and Moglen not only understand the potential pitfalls, but are working very carefully to address them. That's a major reason the new GPL is taking so long to draft.
Time will tell whether the GPL v3 is a success or a huge mistake, but whatever your opinion of RMS and Free Software, there's no reason to believe he and Moglen will make such rookie mistakes that people keep imagining.
I could also argur that I'm the user of the root dns servers, should they be forced to give out the code of there dns-server if thats gpl based?
That depends on how the GPL v3 ends up being written, but would that be a bad requirement? I don't see why.
They *are* creating a feeble version of that requirement. As it is backed by much legalese it is very dangerous and is very likely to cause problems in the future.
Have you seen the GPL v3? You cannot say things like "they *are*" and "is backed by much legalese" without having seen the new GPL (which is quite impossible as it doesn't even exist yet).
You can rest assured that RMS and Moglen understand the issues involved, and are considering them very carefully.
After there are enough incompatible versions of the GPL, how much do you expect to see left of OSS? Licensing gridlock will ensue and the much feared and preached against reinvention of the wheel will be forced. I'd love for you to prove to me wrong on this though.
Prove what wrong? That incompatible licenses create friction? Of course they do. The only alternative is to all stick with one license (or one set of compatible licenses, which really just amounts to the same thing).
This is not as big of a problem you are implying, though. The proof is that we already *have* incompatible GPLs and a whole slew of incompatible Open Source licenses, and yet the Open Source/Free Software ecosystem is very robust.
>Web sites that do this clearly run counter to the intention of the GPL, thus it's a loophole. Why?
Because the authors of the GPL have said so, and they are the ones who, by definition, decide their intentions when creating the GPL.
If I wrote a book using a modified version of a GPL word processor, would I have to publish the modified code?
Nope, and that would not only be a stupid requirement, and would violate the freedoms the FSF intends to promote.
The fact is that web services make the web surfer the user of the program, not the webmaster. Being the user of the code, the FSF wants them to have the rights they've outlined.
The web has created a new class of computer user which the GPL does not adequately address as per the intentions of the authors of the GPL, so they are quite rationally working to update the GPL.
Websites that do this don't clearly run counter to the intention of the GPL as I support it.
That's your opinion. However it's the opinions of the authors of the GPL (RMS and Moglen) who get to define the intention of the GPL. As they have found a flaw in the GPL which allows GPL licensed code to be used in a way counter to their intentions, they are taking responsibility and working to bring the GPL more in line with their intentions.
Expanding the GPL to force source disclosure to anyone who recieves the output of GPL code is absolutely unreasonable.
Believe me, they understand that and understood it long before you ever even considered that possibility, and that's why they don't intend to create that requirement.
If you are embedding, say, GPL md5 checksum calculator in otherwise proprietary software, will you need to publish the whole source code? If so, I'm willing to bet most companies would rather re-invent the wheel and rewrite it. I wouldn't want to publish the source code to a production web site for obvious security reasons.
The algorithm for calculating md5 checksums (and any other crypto or crypto-style processes) are made publicly available. This is because these security features aren't secure because of obfuscation, but because of the rules of mathematics. Releasing the code will only weed out poor crypto[*], not compromise good crypto.
So there are *no* legitimate security concerns.
As for code 'tainting', it all depends on how you incorporate the code, but if a company wants to incorporate GPL'd code, they'll have to abide by the license, which would require, generally, for that specific executable to be released under the GPL. This is by design, and the FSF is fully aware that this design will act as a hinderance to the adoption of their code. While they, I'm sure, want their code used, they will not sacrifice the freedoms they espouse to accomplish those ends. If a company wants to incorporate code without paying for it and without releasing their project, they can seek out and a BSD (or similar) licensed MD5 implementation, as the BSD folks might prefer freedom, they will not sacrifice the ability of others to use their code for those freedoms.
[*]in fact, MD5 has been shown to be weak by this very process, and we are all moving on to stronger hash/checksum systems.
If they want to stick with their own GPL2 fork, they can still keep it locked up.
That's the way it's supposed to work. The reasoning is that the GPL is intended to protect your freedom, so no future GPL is ever allowed to take away any freedom you currently hold.
So Google, for example, will have the right to keep their current code proprietary, even if GPL v3 would disallow that.
That's the "Free Software" aspect of the issue. Now, onto the "Open Source" aspect...
Assuming GPL v3 has the proposed changes, Google (again, for example) will have to either avoid incorporating GPL v3 code into their proprietary GPL software, or abide by the terms of the GPL v3 and make their proprietary source code available.
So there will definitely be pressure for them to release the code when it comes time to incorporate GPL v3 software or merge/upgrade with future releases of their current software, if those projects they rely on move to GPL v3.
Slashdot Mirror
Seems a good anaolgy to me.
That says more about you than it does about the quality of the analogy.
Casablanca is still a great movie, you can watch it "in 2005"; but it's not a movie "of 2005".
But the iPod *is* a product (specifically, a brand with current products) of 2005.
The rules for the Academy Awards are not the same as the rules for PC World's list.
and like movies, sequels rarely win awards.
That's one of the dumbest things I've read all day. Look at the list and notice all the products which are "sequels".
In fact, the iPod that's on the list is a "sequel", and clearly it won.
the point was it's not the original 2001 one,
No one said the original iPod should or even could win. You know why? The original iPod hasn't been an contemporary product for years now.
The dateline of the iPod Photo article: Friday, November 19, 2004.
Your premise, that a product that was not released in 2005 is ineligible for a "Best Products of 2005" award is moronic, and the fact that products that debuted before 2005 (Google is on the list, you idiot) doesn't seem to clue you in.
You're defending a lost position and you know it. Your reasoning is not self-consistent, nor is it even consistent with the article.
Don't bother replying, you'll just make yourself look more foolish. The iPod (in various flavors) *is* a product of 2005, it *is* on the list, and it is *not* a movie. These facts seem fairly basic, and quite easy to grasp.
You can still watch Casablanca at your local revival cinema, so why doesn't it get an Oscar every year?
Why would you think the rules for winning an Academy Award apply to a PC World "Top 100 Best New Products of 2005" list"?
I mean, seriously, are you trying to claim that since Casablanca can't win an Oscar for 2005, that the iPod can't be considered a "Best Product of 2005"? One's got nothing to do with the other.
That's the "new, 2005" video version, I think.
There is (currently) no '"new, 2005" video version'.
You seem to be under the mistaken impression that this is a "Top 100 Best New Products of 2005" list (which it isn't), and not a "Top 100 Best Products of 2005" list (which is what it actually is). Last I checked, the iPod is not only still a product in 2005, but is a very popular and highly acclaimed product.
There *are* better things out there than the iPod. How is this surprising?
It's surprising given that the iPod usually takes a couple of slots in the top ten on lists like these.
And when you have such a generalized list...well, you will always get strange results. What was the criteria for determining a product that would make the list?
Yeah, that's just it. Why would the iPod fare lower than a hard drive? Or a Rio mp3 player?
I mean, if I were going to bet on whether an iPod or a Rio would rate higher on a "best products" list, I'd bet on the iPod. This isn't a earth-shattering surprise or anything, just a "huh, didn't expect that" sort of surprise.
>Especially since it's about products of 2005; the iPod debuted in 2001.
What's that got to do with it? Debuting before this year doesn't mean it's no longer a product.
This is easily demonstrated by the fact that the iPod is on the list.
As are more than a few other products that debuted before 2005, such as Google, and Photoshop Elements.
OK. If it's so obvious that a person is 'using' the backend program, then the program is clearly being distributed to them (otherwise how are they using it?), so I struggle to see how the current GPL fails to cover it.
Distribution is another one of those words which has an objective meaning. Distribution means the person is receiving the binary code, for example on a physical disk, or via download. The web surfer doesn't receive the program, they just operate it.
So, no, the program is not "clearly being distributed to them".
This is similar to how you can use a TV from across the room, without actually touching it. Remotely operating something has been around for quite some time now.
...
You, on the other hand, seem to be wilfully ignoring my opinions,
...
I simply don't believe that the user of a website is any more the user of the backend application that generates it than the reader of a book is the user of the word processor that generated it
The definition of a "user" is not subject to your views, opinions, or beliefs. It's a well-defined, and objective term. The GPL is not contigent upon your beliefs.
It's so simple as to defy belief you can't understand it. *You*, the web surfer, are the one who entered data into the web form and clicked the "Submit" button. That's the interface by which you operate the remote program.
This is analogous typing a command in a terminal window and pressing "enter".
"But," you cry, "the web surfer isn't the one sending the command, it's Firefox or IE!" Are you so simple you are confused by a layer of abstraction? If you load a tiff in the GIMP, the GIMP calls libtiff to process the image. It does this by your request. This makes you a user of libtiff as well, even though you never personally, nor directly, called that library.
With a web browser, you are personally instructing and interacting with a remote program. Which is to say, you are using it. All your views, beliefs and opinions cannot change that fact.
(or users of an application which was compiled using a modified, GPL-based compiler, and if I understand the GPL correctly, in this instance there is currently no compulsion to release the modified compiler).
Right, and the new GPL won't require that either, because running a program that someone else compiled doesn't make you the user of the compiler.
From TFA
"In a broader sense, Arnold believes Google is building a "patent fence around search" technology as the firm moves to codify its unique competitive advantage."
Is this good or bad?
Guns are bad.
Given that guns exist, if the police use guns to fight something worse than guns, it's (potentially) good.
Patens are bad.
Given that patents exist, if a company uses patents for a greater good, it's (potentially) good.
While there is no doubt a contigency in the slashdot crowd that has an irrational need to be absolutely consistent in a superficial sense, there is a valid and consistent logic to being opposed to the existence of patents while, accepting the fact that they do exist, supporting their use for good causes.
I personally don't see anything in the *original* intent of the GPL (not just in the subtleties of the wording) that is being breached by using GPL
The preamble of the GPL begins with, "The licenses for most software are designed to take away your freedom to share and change it. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users." and continues on from there to further clarify the intent behind the GPL.
If you didn't "see" that in the GPL, you clearly haven't even read the GPL.
If Stallman wanted to extend the purpose of the GPL
He doesn't want to extend the purpose of the GPL. As you can see above, he wants to protect the freedoms of "all its [free software's] users". "All" includes those who use software remotely.
My view is that when browsing a website, you are 'using' the HTML and any javascript thats running.
Nice try, but it's not the 'using' of the HTML output that is in question. It's the issuing of commands (such as via HTTP), which makes you the user of the remote software. The data sent via HTTP (such as HTML, JavaScript, images, and so on) is the output of the software on the remote server that you, the web surfer, are using.
No. I'm flabbergasted that anyone can see this as tightening a loophole as opposed to a significant change to the scope of the GPL. As I think I've already mentioned Stallman is free to do what he wants in creating a new GPL, but trying to claim it's simply tightening a loophole is just plain wrong.
I've never said it wasn't changing the scope of the GPL. In fact, I've stated that that's exactly what it is *many times*.
You can't read English, can you? The loophole isn't between the GPL and the usage in question. It's between the intention behind the creation of the GPL and the use in question.
Quit being so dense.
Read carefully (I know you won't, but I type this anyway):
1. Stallman wishes to promote four freedoms.
2. He creates the GPL with that goal in mind.
3. The GPL does this splendidly.
4. A relatively obscure, yet absolutely valid, usage of the GPL becomes common which allows people to violate the four freedoms.
That's the loophole. It's between 2 and 4, not between 3 and 4. I've pointed this out about a million different ways. If you can't see that, and you can't stop parroting your original, "but it doesn't violate the wording of the GPL", then my participation in this thread is over.
The whole reason loopholes exist is because people are able to find ways to violate the intent of a thing (a law, for example), while complying with it to the letter.
What is the 'browsor' (my term for a browser operator) using, other than his browser?
It's really simple. The user is the one who uses the program. For interactive programs, the user is whoever is clicking the buttons, typing the keys, etc.
If you are using a program which in turn sends commands to another program, *you* are using the second program as well.
For example, if you were to put a program online which creates images of the current sky (astronomy/planetarium software), and the user is presented with a form to enter their coordinates, and a few other parameters, the web user is, in fact, using their web browser to control the astronomy program. In other words, they are using the astronomy program.
The same applies with Google. If Google is using modified GPL'd code to dynamically serve up maps, the user who directs their browser is using that GPL'd code.
The person in question is not the 'user' who is running Apache, perl, Python, Ruby, or whatever on the server. That person is 'anonymous', 'guest', or 'nobody' on the server.
That is a user id. A user id is not a person, it's a number. It's nonsense to say "anonymous" or "guest" or even "billg" is the user. The user is the actual person using the program (and who may or may not be using some given user id or another). If you are logged into your X11 session as "themonster", and you have the GIMP running, if you hand me the mouse and keyboard to do some photo editing, I am the user of the GIMP, not you.
Now, of course, licensors are free to ignore this distinction, and create a per cpu license, or per instance license, or whatever. But the GPL is a license intended to provide a certain set of freedoms to the user, so it's logical to apply this to users of remote software as well.
For the millionth time, I fully agree that there's no loophole between the wording of the GPL and the use that's under question. You can quit saying things like, "What I'm saying is that I see nothing in the original GPL that suggests there was ever any intent to cover this sort of scenario."
Read carefully: The intent of the GPL is to protect four very specific freedoms. The GPL, as worded, fails to do so. It's a way for people to comply with the letter of the contract while violating its intent--which is to protect the four freedoms. Please don't respond with, "but the use in question is not covered by the GPL," because I've already made it abundantly clear that I agree it isn't.
The point I'm trying to get across is that 'web-enabling' a system doesn't seem to me to be anything particularly special (and often doesn't involve doing anything specifically to the app at all). People have been interacting with companies' back-end systems through a variety of direct and indirect means for years, and none of these have been seen as 'distributing' the app (for the purposes of the GPL), and I don't see why a website is any different.
That's because you are looking at the issue in the wrong way. RMS isn't primarily concerned about distribution, he's primarily concerned about freedom. The distribution aspect of the GPL is the means by which he's attempting to legally protect the four freedoms he believes should be protected.
When you look at it simply from the point of view of distribution, you are correct that the GPL doesn't really apply to web services, and you know what? RMS agrees with you. That's why he's working to update the GPL. Absolutely everyone is saying that the GPL doesn't apply to web services. Google is saying it, RMS is saying it. I'm saying it. Yet still, you jump up and down screaming, "but it doesn't apply to web services!!!" as though you're arguing with someone.
Basically, I'm struggling to understand how (or why) the licence requirements for an application can change the second the app becomes involved in a web transaction,
That's up to the license. Most licenses don't make that distinction, such as the current GPL. Other licenses do, such as the license for IIS, Exchange, SQL Server, etc.
To me this feels a fundamentally different act to that of releasing an application binary for someone else to pick up and install.
That's because it is a fundamentally different act. No one is claiming otherwise. The commonality is that the four freedoms in question can apply to both situations.
You're right that people are free to choose whether they will use GPL3 code in the future, but it is a big deal. I know that the next enterprise app I design will not include any GPL3 code if it means that as soon as I web-enable it, I'm forced to publish my private business logic(whereas I've got no problem with the current GPL as I've got no intention of making my apps available to third parties to install).
Assuming you do hardcode your business logic, and assuming you do wish to keep it secret, then you are fully free to do so. The purpose of the GPL isn't to make free software palatable to someone who wishes to keep their software secret. The GPL is designed with the primary intention to allow for people to create software which provides the user with four specific freedoms.
You're confusing the intention of the Open Source community with the intention of the Free Software Foundation. They have two very different goals. To the Open Source community, it *is* a big deal if their license makes it hard for you to use their software. To the FSF, the big deal is when you are allowed to supress the four freedoms they wish to promote.
It all depends on what you want. You sound like your are more of an Open Source kinda guy, so it's no big deal that you don't like the GPL 3, since you aren't the target audience.
You seem flabbergasted that the FSF would want to update the GPL to cover web services, but if you understand their intention, it's quite reasonable.
So where am I getting at? The key reason why Microsoft got in trouble was it's INTEGRATION (IE as a part of the OS) and LACK OF CHOICE (media player installed by default) and not the fact that it was bundled on the CD.
The problem wasn't the integration so much as the fact that MS, by virtue of providing the OS used by over 90% of all PCs, was able to (and in fact did) use that product to force other products on the user. The problem *was* the bundling.
In retrospect, the bundling issue sounds incredibly lame. What modern OS doesn't include a media player or a web browser as part of the/a default install?
It's silly. OS X bundles Safari and QuickTime. GNOME bundles Totem and Galeon. KDE has Konqueror and whatever their media player is. Virtually all distributions bundle Firefox and xmms. Even BeOS and Amiga OS have bundled browsers and media players.
MS bundles IE, but I used Netscape exclusively until IE 4, and I now use Firefox. MS bundles Media Player, but I still use VLC, QuickTime, Real, iTunes, and about a million other media players.
My problem is, as I mentioned, that as far as I can tell, the GPL talks about the distribution of the software, not the use of it.
That's correct. I've already said that. The loophole is not between the wording of the GPL and the use of the GPL, those two things are in perfect sync, as regards to web services. The loophole exists between the intention behind the GPL, which is not about distributions, but about promoting four very specific freedoms.
I'll state this very clearly:
You are correct that there is no loophole being exploited related to the wording of the GPL.
There is, however, a loophole related to the intent of the GPL.
Quit bringing up the wording. We agree on that aspect, and so does Stallman. If he didn't agree with that, he wouldn't be so set on changing the GPL.
Say I've got a (non-web enabled) account management system, which includes some GPL code, and some proprietary business rules. My understanding is that under GPL I would not have to publish the source code for this app.
Correct. And "proprietary business rules" sounds like your talking about configuration of a compiled binary (similar to rules in a mail app). You've not changed the source code, but for the rest of this example, I'll assume you have.
Now say I build an website on top of it, allowing the users to view and update various aspects of their accounts (such as changing their address). At this point, would I have to publish the source code for the backend account management system?
Under the current GPL, no. Under the new GPL, assuming what we hear of it actually comes to pass, yes. That's exactly what the authors of the GPL intend.
Stallman believes users of software should have four specific freedoms, and he has written a license specifically designed to promote and protect those four freedoms.
I don't know if you posed this scenario because you didn't know the answer, or if you intended to show a scenario which you find undesirable (it can be read both ways). But if it's the latter, I don't see why someone would really be bothered by the fact that they have to provide access to the source code of software they got entirely for free. But as with all licenses, the author of the license (RMS and Moglen, in this case) are free to include any legally allowable terms they desire, and you are equally free to accept or decline those terms.
This is not a big deal, and if you are currently using software licensed under the GPL, you will not have to abide by version 3 of the GPL if you don't want to. In other words, if you are currently exploiting the loophole in question, you will be allowed to do so forever. The new terms only come into play if you accept software under the GPL v3 (or later) license.
That's very different to a "loophole" where the intention of the licence is extremely clear but people are using clever tricks to avoid it.
The intention of the license is very clear--it's to promote four very specific freedoms. In the opinions of RMS and Moglen, the current version of the GPL fails to ensure those freedoms in a specific set of circumstances. This is due to a loophole--a loophole not between the wording of the GPL and this specific set of circumstances, but a loophole between the intention of the GPL and that set of circumstances.
To my knowledge, the current GPL talks about copying and distributing, not using.
Copying is where the GPL applies, because it's based on copyright law. Distribution is where they chose to focus because:
1. This keeps the GPL simple (too complex and unintended consequences are more likely)
2. Distribution was the most prevalent way users became users.
Now, it's increasingly common for a person to become a user (as per the opinions of RMS and Moglen, and their opinion is what applies regarding the intent of the GPL, as it is their document) without having the software they use having been distributed to them. Therefore, in order for the GPL to more accurately address the intentions of its authors, it will need to be updated, which is exactly what they are in the process of doing now.
Indeed it would - so if I used my program to write static HTML which I then published, would this program require me to publish the source?
I'm fairly certain they do not intend this, and that the new GPL will not require this. If you read the pages on the FSF site, it will become clear that they believe users should have the right to use modified GPL'd software without distributing their changes if they so desire. Your static HTML example would fall into this category.
It's only because web surfers are, in effect, using the software that this question has come up.
If you use a program to create a static html page, you are the user. If I click a button on a web page, the program that is run by my click is now being used by me, and thus I am the user.
At least, as far as RMS is concerned, as I understand it.
Yes, by limiting the freedom of people to use free software for purposes Stallman doesn't like.
... that's bad for the free software world, because if it happens too often, people will standardize on something else and free software will become marginalized.
Absolutely wrong. Stallman doesn't like Communism, but his license doesn't say the Chinese can't use emacs to draw up the death-sentence for some dissident.
The *sole* purpose of the GPL is to promote the FSF's Four Freedoms. Nothing more, nothing less.
It's an irony, but inescapable fact, that to promote freedom, one must actually limit freedom. It sounds illogical, but it's not a matter of logic, it's a matter of reality. For example, if you don't have laws that limit someone's ability to kill someone else, you have less overall freedom because the threat of being killed can be used to force people to do what you want.
Similarly, the GPL limits certain freedoms (more specifically, it outlines certain responsibilities, and limits your ability to add further licensing restrictions to a piece of software) in order to promote more freedom overall. Of course, just as in the real world, it's possible to go overboard and limit freedom in such a way that is not rational. I'm sure you can think of some laws right now that do that. The GPL, on the other hand, does not do that. The GPL is simple, is predicated on four basic freedoms, and is fully consistent with those four freedoms.
Plus, let's not forget what happens when you make free software too difficult for commercial entities to use
RMS's goal is not for all software to free. His goal is for free software to exist. He has stated many times that you are perfectly free to use non-free software all you want. ESR is the one who's more concerned with promoting a system he sees as superior for practical reasons (read The Cathedral and the Bazaar), and while he's a libertarian, he really isn't primarily focused on promoting freedom, he sees that as a nice side-effect.
So, stating that the new GPL may make some companies not use the GPL is not a counter argument because it doesn't have "companies use GPL software" as its primary goal. If you want that, the BSD license is more in line with that goal. You (and everyone else) is free to chose whatever license they wish for their own fully-owned code.
When he says "working", you and I both understand that he means "working insofar as allowed by necessary security restrictions put in place to prevent abuse of the service." But the GPL is a legal document, and unless that's explicitly stated, it could be interpreted as meaning no restrictions allowed, period. And that's how corporate lawyers and suits will interpret it - encouraged, no doubt, by the various FUD factories.
It's a bit premature to parse words that aren't the actual license. The GPL currently requires one provide access to the source code via a network service (like ftp) or on physical media, and no one has argued that the ftp site has to allow itself to be overburdened, or that a person has to ship out CD's immediately. What's likely to be the case is that the idea of "reasonable" will be applied. In other words, what does the industry consider to be "reasonable" access via ftp? Or a "reasonable" rate of delivery of physical goods?
Maybe it'll be used as FUD, but I really doubt it, as I've never heard (until you brought it up just now) anyone imply the GPL would require you to overtax yourself or your equipment in order to provide access to source code.
I think you're wrong. The user of a program is the one that downloads installs/authorizes code execution on their system.
Your definition of "user" is flawed. A user is (it's in the actual word) the one who uses the program. That's it. One doesn't need to download, install, or authorize a program to be a user, one merely needs to use the program. That often entails downloading and such, but doesn't require it.
The owners of a website use a web server and web apps to deliver output to the persons browsing the site.
No, they don't. A computer delivers the output. The owner of the website provides/allows access to the services, the services provide the output.
Likening use of HTTP to linking a library is wrong, and IMO idiotic.
Good thing no one is doing that.
I find the uproar over this quite amusing. If MS, Apple, IBM, etc, can restrict network use of their software via license, why can't the FSF? The fact is, they certainly can.
There are really only two questions:
1. Is this in line with the ideals of the FSF? I believe it is, but more importantly, RMS thinks it is.
2. Will this cause unacceptable problems? It's up to those who are in a position to accept the license to determine whether it's acceptable or not, which requires an actual license to begin with. As for whether it causes problems is up to RMS's and Moglen's ability to draft a useful license. They have shown proficiency at this in the past, so it's a bit unfair to assume the worst.
Not in the sense you're talking about, of course not. But, if you're legally bound to leave a particular service open and unthrottled, it could become a potential vector for DDoS attacks. There need to be provisions to allow a site admin to place limits on the number of simultaneous requests, the frequency of requests from a single client, and other such limits to help prevent abuse.
That's a very odd thing to say, and I don't quite understand where you go the idea that webmasters or site admins would be required to leave a service open and unthrottled. In fact, that would go counter to the FSF's "four freedoms".
It will be backed by legalese (though a human friendly dialect where possible), since last I understood the nature of a license is legal, and they want to be legally clear and unambigous about their license. Again, feel free to prove otherwise.
Prove what otherwise? That the GPL is a legal document? Don't be silly. That's not what you implied. You implied that it will be "very dangerous" because it "is backed by much legalese". Without having seen it, you aren't in a position to judge how "much" legalese it will require, and how "very dangerous" that will make it. I can't disprove your fears, but it's not my requirement because I'm not making the claim, you are. My claim is merely that you don't have the necessary facts on which to found your statement, which I can prove by the simple fact that the GPL v3 doesn't even *exist* yet.
The problem is that making more licenses for shits and giggles is a bad idea
The GPL v3 isn't being created for "shits and giggles", it's being created to address a very real concern.
and over time will be an extra burden at the very least.
I agree, which is what I meant when I wrote, "that incompatible licenses create friction? Of course they do." I thought that was fairly clear.
You bring up valid concerns, but you can be very certain that RMS and Moglen not only understand the potential pitfalls, but are working very carefully to address them. That's a major reason the new GPL is taking so long to draft.
Time will tell whether the GPL v3 is a success or a huge mistake, but whatever your opinion of RMS and Free Software, there's no reason to believe he and Moglen will make such rookie mistakes that people keep imagining.
I could also argur that I'm the user of the root dns servers, should they be forced to give out the code of there dns-server if thats gpl based?
That depends on how the GPL v3 ends up being written, but would that be a bad requirement? I don't see why.
They *are* creating a feeble version of that requirement. As it is backed by much legalese it is very dangerous and is very likely to cause problems in the future.
Have you seen the GPL v3? You cannot say things like "they *are*" and "is backed by much legalese" without having seen the new GPL (which is quite impossible as it doesn't even exist yet).
You can rest assured that RMS and Moglen understand the issues involved, and are considering them very carefully.
After there are enough incompatible versions of the GPL, how much do you expect to see left of OSS? Licensing gridlock will ensue and the much feared and preached against reinvention of the wheel will be forced. I'd love for you to prove to me wrong on this though.
Prove what wrong? That incompatible licenses create friction? Of course they do. The only alternative is to all stick with one license (or one set of compatible licenses, which really just amounts to the same thing).
This is not as big of a problem you are implying, though. The proof is that we already *have* incompatible GPLs and a whole slew of incompatible Open Source licenses, and yet the Open Source/Free Software ecosystem is very robust.
>Web sites that do this clearly run counter to the intention of the GPL, thus it's a loophole.
Why?
Because the authors of the GPL have said so, and they are the ones who, by definition, decide their intentions when creating the GPL.
If I wrote a book using a modified version of a GPL word processor, would I have to publish the modified code?
Nope, and that would not only be a stupid requirement, and would violate the freedoms the FSF intends to promote.
The fact is that web services make the web surfer the user of the program, not the webmaster. Being the user of the code, the FSF wants them to have the rights they've outlined.
The web has created a new class of computer user which the GPL does not adequately address as per the intentions of the authors of the GPL, so they are quite rationally working to update the GPL.
Websites that do this don't clearly run counter to the intention of the GPL as I support it.
That's your opinion. However it's the opinions of the authors of the GPL (RMS and Moglen) who get to define the intention of the GPL. As they have found a flaw in the GPL which allows GPL licensed code to be used in a way counter to their intentions, they are taking responsibility and working to bring the GPL more in line with their intentions.
Expanding the GPL to force source disclosure to anyone who recieves the output of GPL code is absolutely unreasonable.
Believe me, they understand that and understood it long before you ever even considered that possibility, and that's why they don't intend to create that requirement.
If you are embedding, say, GPL md5 checksum calculator in otherwise proprietary software, will you need to publish the whole source code? If so, I'm willing to bet most companies would rather re-invent the wheel and rewrite it. I wouldn't want to publish the source code to a production web site for obvious security reasons.
The algorithm for calculating md5 checksums (and any other crypto or crypto-style processes) are made publicly available. This is because these security features aren't secure because of obfuscation, but because of the rules of mathematics. Releasing the code will only weed out poor crypto[*], not compromise good crypto.
So there are *no* legitimate security concerns.
As for code 'tainting', it all depends on how you incorporate the code, but if a company wants to incorporate GPL'd code, they'll have to abide by the license, which would require, generally, for that specific executable to be released under the GPL. This is by design, and the FSF is fully aware that this design will act as a hinderance to the adoption of their code. While they, I'm sure, want their code used, they will not sacrifice the freedoms they espouse to accomplish those ends. If a company wants to incorporate code without paying for it and without releasing their project, they can seek out and a BSD (or similar) licensed MD5 implementation, as the BSD folks might prefer freedom, they will not sacrifice the ability of others to use their code for those freedoms.
[*]in fact, MD5 has been shown to be weak by this very process, and we are all moving on to stronger hash/checksum systems.
If they want to stick with their own GPL2 fork, they can still keep it locked up.
That's the way it's supposed to work. The reasoning is that the GPL is intended to protect your freedom, so no future GPL is ever allowed to take away any freedom you currently hold.
So Google, for example, will have the right to keep their current code proprietary, even if GPL v3 would disallow that.
That's the "Free Software" aspect of the issue. Now, onto the "Open Source" aspect...
Assuming GPL v3 has the proposed changes, Google (again, for example) will have to either avoid incorporating GPL v3 code into their proprietary GPL software, or abide by the terms of the GPL v3 and make their proprietary source code available.
So there will definitely be pressure for them to release the code when it comes time to incorporate GPL v3 software or merge/upgrade with future releases of their current software, if those projects they rely on move to GPL v3.