Number of GPL v3 projects tops 2,000

Da Massive writes "The number of open-source projects that use the GNU General Public License Version 3 has grown to more than 2,000, according to Palamida, which sells software and services for tracking open-source code within a customer's code base. 'Our database now contains over 2,000 projects that are using the GPL v3. "At this rate the GPL v3 is being adopted by 1,000 projects every 4-5 months, and if the trend continues, the license will be used by 5,000 projects by the end of the year," states a recent posting on Palamida's blog.'"

Calm Down

[stry_cat]

I'd like to know how many of the projects fall into the "Hey look my project which no one other than myself contributes to and uses, is now using GPL3" Until the kernel switches from GPLv2 it won't really be considered a success in many people's eyes.

However I have a more important question. Why is this written like it is a war between GPLv3 and other licenses? If the virus of GPLv3 doesn't spread we're all doomed. No folks that's not the case. Don't get so wrapped up in this stuff. So what no one uses GPLv3? So what if everyone uses it? If software doesn't meet one's needs (and that includes having cumbersome provisions in your license) one will either write their own or use someone else's software. Really this all works out in the end. Don't have so much emotion invested in things that you can't really affect the outcome.

Re:Linear interpolation...

[Otter]

It looks like they're adding ~180 new projects per month. With nine months remaining, that doesn't come out to 4000 by year end, let alone 5000.

Additional thoughts:

1) Any time I read something involving "If this trend continues", even if it's based on solid data I hear it in my head in Disco Stu's voice, which tends to undercut its credibility.

2) This ("Four new GPL 3 projects this week!") is arguably the most boring blog in the world.

Re:Twice nothing is still nothing ...

[qortra]

How many GPLv2 projects are there out there? Easily over 100,000

Out of curiousity, where are you getting this statistic? It might be more accurate, but on the other hand, it is probably more reasonable to compare numbers in the same database (with consistent identification methods and all).

According to the blog [blogspot.com], there are "6446 GPL v2 or later" projects in the database, and assuming that more than 2000 of those are v3, that leaves just a little bit over 4400 are v2. Assuming the databased isn't biased towards v3 projects in any way, that means that almost 30% of projects are v3.

I'm not necessarily saying that your statistic is wrong; I would just like to see a source, and I definitely would prefer to do apples-to-apples comparisons (consistent collection methods).

Re:GPLv2 compliance-?

[morgan_greywolf]

This is exactly what's stopping my project from adopting GPLv3. Stylus Toolbox uses GladeGen [linuxjournal.com] for some of its more complicated windows (I did the initial design using GladeGen, but the rest of the project just uses stock PyGTK code).

The problem is that the linked-to article, all the documentation that exists for GladeGen, only mentions that the code is GPL; it doesn't specify a version and there is no COPYING file or mention of a license in any of the files or source code. So I'm not entirely certain as to the legal status of the code other than it's probably licensed under some version of the GPL.

If David Reed or Linux Journal could come forward and clarify the legal status, that would be most helpful. Maybe someone with a legal background might provide some help, too.

new projects??

[gnurag]

Is that number of new projects started with GPLv3 license or projects relicensed under GPLv3?

Re:Twice nothing is still nothing ...

[qortra]

For those of us who have moved completely away from Microsoft, [Samba] IS nothing ...

False. Samba is a fantastic utility, and a very acceptable protocol (it would be great if it were completely open). It is mountable in a unix filesystem, it can be safely authenticated over untrusted networks, it is spoken by countless devices and operating systems (not just Microsoft), supports two way transfers, supports simultaneous transfers, and it is reasonably fast. You stated 4 replacement protocols. Here is an analysis of where and why they fail my criteria set.

• svn - its none-trivial usage aside, it is not designed for the same kinds of transfers as Samba. Primarily, it isn't designed for streaming. For instance, I wouldn't consider it viable for the archival of playback of music and movie collections.
• ftp - stop joking with me. Usually single threaded transfer, laggy, not usually viable as a mounted disk (unless Fuse is used), and unsafe authentication in untrusted environments. This protocol is ancient, and it shows its age badly.
• http - by default, unidirectional. Requires specific and non-standardized handlers to send information the other way. Also, no standard for file permissions, no safe method for authentication (without https).
• ssh - The only one in this bunch that actually has a shot. It meets most of my requirements, especially now that you can mount an ssh filesystem with Fuse. However, that is still a fringe mechanism for mounting. Most embedded devices (even Linux based ones) do not support this. They support Samba. For examples (in case they are demanded), look at Media and Content streamers at AVS Forum [slashdot.org]. Many of them are Linux based, but almost all of them support Samba, uPNP, and possibly Apple Rendezvous (or whatever it is). None of your protocols can be found.
• nfs (which you forgot) - Also close. I spent a lot of time trying to make NFS work in my network, but much to my chagrin, it simply isn't safe over an untrusted network. You can always spoof file UIDs, file GIDs, and IP addresses (unless you have really nice switches).

Re:And this matters, why?

[cbart387]

Agreed. I have a problem with the viral aspect of it since you're telling people it must be wide-and-open. If I ever start a project I'll keep it open because I believe in sharing. I do not believe in telling others to share because that's for them to decide. I'll stick with the BSD or even better the beerware [wikipedia.org] license. In my definition of free, they're more free than the GPL, despite RMS's opinion.