In an October 9, 2007 article on Vnunet regarding GPL v3 conversions, the author refers to Black Duck's CEO in the piece and writes, '[He]suggested that most projects that have converted to using GPLv3 so far are the smaller ones.'
We find this to be a very interesting assessment of what is going on in our community and within projects that may not consider themselves small. Ernie Park, Palamida's Director of Research points to Palamida's publicly available database as a true benchmark for GPL v3 conversion activity and states, "Our database tracks all conversions, and all projects influenced by GPL3."
Palamida's opinion is that developers want, and in many cases, need to know if the open source they are currently embedding, or considering using, is going to migrate to the newer license. An "intent to migrate" can spell serious issues for anyone in the embedded market currently reliant on GPl v2. If their particular project of choice decides to convert with the next release, they have serious legal and business issues on their hands. Thus, keeping track of the ongoing conversations is a critical part of proactive open source code management.
Park explains, "Our conservative numbers are far more accurate than any guess-timates. Only 883 projects have converted? Smaller projects only? I guess nobody heard of Rubyforge and Savannah, the repo for GNU, Samba, or Sugar CRM, all big adopters, and core to much of what is done in open source." There are of course the FSF supporters who moved over within days of the GPLv3 release, but many derivatives of these core projects exist across the board, up and down the stack within the enterprise environment.
Park says that in order to get an accurate view of what's happening with GPL v3 conversions, people need to consider the following:
31,000 active open source projects as calculated on the GPL v3 blog
5500 "or later"
800 GPL v3
More than 1% converted each month so far.
"or later" is important because while these licenses do not state explicitly that they have moved over, they are compatible with the GPL v3 and do not discourage its use. Ignoring these license terms when considering overall conversions or acceptance, is a huge oversight.
So, with more than 20% of the active open source software in use potentially encumbered by the "or later", along with a conversion rate of 1% per month, currently representing 2.5% of the active projects, Palamida feels that the conversion is significant.
What does this really mean?
With 20% of OSS in use encumbered by "or later", and 2.5% of the existing active OSS already releasing versions under GPL v3, if anyone believes that there is not that much GPL v3 in OSS, they just aren't looking. With major GPL v3 releases coupled with the "little projects" making up the nuts and bolts of the Internet, not only is it already there, the coverage is growing steadily.
There has been much ado about the infamous Evans Data report which cites a percentage of developers who state that they will not be using the GPLv3.
Palamida was very interested in their findings as it was never made clear the context of their survey. On the one hand, we would agree with the outcome of the survey, based on feedback we are hearing from development teams at our enterprise customers. Most developers who solely participate in commercial software development would never think about adopting the GPL v3 license.
So a couple questions come to mind - what kind of developers did they poll? What are their roles and within what size/type organization (software companies, F500 companies, etc?). If the survey consisted of just developers who participated in some way in open source projects (whether personal or professional), then we would be surprised at the numbers. For us, this hypothetical point confirms what we hear anecdotally from our customers - most commercial software vendors are reluctant to adopt GPL v3. But of course, this market segment was reluctant to adopt GPL v2 as well...
When looking at the overall landscape of GPL v3 conversions in relation to current internal use policies, legal issues and/or compatibility, Palamida believes that not keeping an eye on the pulse of future migrations can spell trouble for organizations currently embedding open source into their code base. What is GPLv2 today, may easily be GPLv3 two months from now. If you don't know what you've got, how will you avoid the potential legal and business risks associated with the different license types?
I'm just sayin'.