Definitely not. Consider the case of Linksys' WRT-54G. Initially, it ran Linux, but Linksys didn't feel like releasing the source code it was using. This made the device essentially useless (except as an average WAP). However, when they were called on their GPL infringement, they released the code, creating a great community around the WRT-54G. Now the thing can be an OpenVPN endpoint, or a backup MX for quick mail server reboots, or anything you can imagine. If Linux was BSD-licenced, the WRT-54G would probably still be a closed device, which would be bad for the computing community in general. (I think Linksys got a lot of sales because of the openness, and they certainly didn't lose any. But it's hard to convince big companies of this unless their lawyers are forced to be involved -- and the GPL ensured that this happened.)
Also, I like the GPL, "useful" or not. If you want to use my code, I want the changes you make to it. It's only fair.
It's not a loss to me if you refuse to use my code because it's GPL'd. With the GPL, you didn't use it and hence didn't give anything back to me. Without the GPL, you would have used it and not given anything back. So the situation is the same for me either way. (Sure, it's worse for you, but I code with my own interests in mind.)
Your anecdote about linksys does not undermine ESR's central thesis, at all. He argues that under the presumption open-source is more efficient, the market will punish those who remained closed. That is not refuted by pointing out a singular example of a company and their dalliances with various licenses. I think that point is an interesting one to consider and needs a far more well formulated response than simple anecdotes.
Your second point is more in line with the story, namely that the goal isn't "efficiency" but "freedom". People who release their code as GPL'd code for freedom and other philosophical reasons are fundamentally different than those who release it for pragmatic reasons (that, in the end, open-sourced code is better code).
> It's not a loss to me if you refuse to use my code because it's GPL'd.... So the situation is the same for me either way.
ESR addresses this directly. Yes, it is a loss to you. Just not maybe to the function ("fairness") that you happen to be optimizing. If your goal was to create the best XYZ library, you've lost mindshare, potential contributors, and possibly created a competitor. This is true whether they are legitimately scared or whether they are scared due to FUD.
> Without the GPL, you would have used it and not given anything back.
This is not necessarily true. I'll give you three examples.
1. I could want to use your library as a monolithic piece and give back the changes I made to that piece. You haven't allowed that. (ie, versus an LGPL license).
2. I could succumb to FUD and just not want to go down that path, despite the fact that I would have been an otherwise good member of the community.
3. I could enter into the relationship intending to withhold my changes, eventually become dependent on having the best XYZ library in the world, and eventually realize the most cost-efficient way to do that is to share my changes and have them be incorporated into everyone else's (which we are presuming is true, since we are presuming open source > closed source).
In the ways that the article mentions, the GPL is as useful as it ever was. Nothing has changed that changes the merits of the GPL. It still gives you the freedoms that people want in open source software and it still guards against other companies building proprietary things on top of GPL software. If those protections are what you're looking for, the GPL still provides them as it always has and the Apache license doesn't offer you those protections.
I'm not saying whether those protections are good, just that the Apache license and the GPL haven't really changed their stances in a way that one is just so much more 2009 than the other.
However, the GPL is outliving its usefulness in another area not mentioned by the article. The GPL is great for software that gets distributed - like a music player that you download or a web browser or whatnot. Once someone distributes it, they have to abide by the license's terms. However, more and more our software isn't being distributed. We're using cloud software which isn't distributed to us. As such, even if someone is using modified GPL code, we have no standing to request the code from them since it wasn't distributed to us in binary form.
You might not agree with the free-software philosophy and that's fine, but it is important to see how cloud computing is making it increasingly easy to turn GPL software proprietary. As long as you don't distribute it, you can be as closed source as you want with GPL'd code.
A silly argument. For every argument about how the GPL retards open source adoption, there's a counterargument for how the GPL enables open source authorship. Maybe as much as a plurality of all corporate-owned/sponsored open source software would never have been released if the only way to do so was to allow your competitors to shrink-wrap it and sell it against you.
I see this as less of an intellectually honest argument against the GPL than it is another futile bid for Eric Raymond to regain the spotlight by picking a fight with RMS. But RMS has already won; the industry embraces and defends the GPL now even as Stallman becomes more and more marginalized and caricatured. There's no room for Raymond in the picture anymore.
Why not combine the 4 rights of the GPL, so dear to RMS's heart (the right to view the source code, modify it and redistribute it) with the obligation to pay the developers a small price.
This would create the best of both worlds: geeks can tweak and modify, consumers pay for the product, so software developers can sustain themselves and receive income for writing great code.
In effect, the license would make the software libre, but not free.
For those who know me, more specifically I'm talking about combining the Software Bill of Rights, which handles the question of how to share the proceeds among a group of developers, with some form of GPL or Apache license.
Definitely not. Consider the case of Linksys' WRT-54G. Initially, it ran Linux, but Linksys didn't feel like releasing the source code it was using. This made the device essentially useless (except as an average WAP). However, when they were called on their GPL infringement, they released the code, creating a great community around the WRT-54G. Now the thing can be an OpenVPN endpoint, or a backup MX for quick mail server reboots, or anything you can imagine. If Linux was BSD-licenced, the WRT-54G would probably still be a closed device, which would be bad for the computing community in general. (I think Linksys got a lot of sales because of the openness, and they certainly didn't lose any. But it's hard to convince big companies of this unless their lawyers are forced to be involved -- and the GPL ensured that this happened.)
Also, I like the GPL, "useful" or not. If you want to use my code, I want the changes you make to it. It's only fair.
It's not a loss to me if you refuse to use my code because it's GPL'd. With the GPL, you didn't use it and hence didn't give anything back to me. Without the GPL, you would have used it and not given anything back. So the situation is the same for me either way. (Sure, it's worse for you, but I code with my own interests in mind.)