The GPL for better or worse (and I believe mostly for the better) is associated with Richard Stallman and it is not unreasonable to wonder if outside of enclaves like MIT (where Stallman resides) and without the spare time hobbyist pursuits of supergeeks, a viable model can exist for sustainable free software development as envisaged by GNU and the GPL. If talented coders are spending time developing complicated applications for free, how do they put food on the table? Is it not by leveraging their skills to either gain an income elsewhere (i.e., a giant closed-source non-free corporation) or gain research, academic or equivalent funding for pursuing such projects? How well do such models even scale?

Stallman addresses these very issues in his essay Why Software Should Be Free. His response is along two lines. The first is the simpler one: as in the case of my Ewing analogy, the issue is really not one of putting food on the table but the pursuit of enormous wealth, and once the wealth factor is discounted, it will be found that such development is not so costly an affair after all. The second line is the more interesting and important one: Stallman points out that software is a public good and is as such best funded using public funds.

Along the way RMS also outlines why, as his title notes, software should be free.

Open Source proponents, in particular Linus Torvalds and Eric Raymond, have argued for a “pragmatic” approach that Jalkut (if I read him right) echoes. If corporations can use the code, thus increasing its usage, why not? While I am not certain that they make such an argument (it has been a while since read Raymond), one can be made that in turn such corporations (e.g: Google, RedHat) can be patrons of open source development, thus solving the funding problem. And therein lies the problem. There are differences between patronage and libre and one of them is that while the latter might require scaling down one’s expectation of wealth (as outlined by Stallman) the former is the one whose sustainability rests on the charity (to put it a bit bluntly) of others.

An aside: do they get it?

In papers such as An empirical analysis of economic returns to open source participation, a strange sort of hand-wringing occurs as a result of the seeming paradox of why someone might write software for free. That the paradox may be a result of the theoretical commitments of the researcher is best left to Amartya Sen to explain (sorry, pay site). RMS explains it in simpler terms in his essay as well, in terms that should be eminently sensible to anyone but a Cartesian sceptic!

An apology: this post was meant to be a much more long-winded one that attempted to explain and defend my thoughts on this issue in a more rigourous manner. But perhaps you are relieved rather than in need of an apology?

1. I’ve never encountered a serious client who chose not to use WordPress because it was GPL-licensed, and I think it’s hard to argue that WordPress’s license has had a dampening effect on its adoption, given its success over competitors with widely varying licenses.
2. I think we have an incredibly strong third-party extension, plugin, and theme community that has flourished, not in spite of the GPL license, but because of it.
3. I’ve seen the absence of GPL in practice; there have been times in the WordPress world when parts of the community have “gone dark” and claimed their code was under more restrictive licenses, like used to be common with themes. Every time this cycle starts it basically kills innovation in that part of the WordPress world until people start opening up their code again or until a GPL equivalent is available. I’ve seen this firsthand several times now.

Gruber then contests the relevance of Matt’s argument (above) — “none of these three points from Mullenweg address Jalkut’s argument”:

1. Jalkut wasn’t arguing about whether users will not use GPL software; his argument was about developers.
2. Jalkut never argued that WordPress wasn’t popular or didn’t have a strong extension/plugin/theme community. Jalkut’s argument was that WordPress might have an even stronger extension/plugin/theme community if it were licensed under a BSD-style license.
3. Jalkut wasn’t arguing in favor of more restrictive licenses; he was arguing in favor of less restrictive ones: BSD/MIT/Apache style ones.

Below I examine each of the points raised by Gruber:

1. Jalkut was arguing about developers, not users

While it is true that Matt touches upon what is good for users, he does so exactly to highlight, inter alia, that what matters is what is [ultimately, if you wish] good for the user. However, as the very sections quoted by Gruber demonstrate, that is not the limit of Matt’s response. Here is Gruber quoting Matt: “I think we have an incredibly strong third-party extension, plugin, and theme community that has flourished, not in spite of the GPL license, but because of it” (emphasis mine). These are developers that Matt is speaking of, not users. Gruber, and Jalkut, disagree with Matt’s conclusion that the community has flourished because of the GPL (that is, of course, the substantive issue), but Gruber’s criticism that Jalkut was arguing about developers but Matt is not, clearly is wide off the mark.

2. Jalkut’s argument is that WP would have a stronger extension/plugin community if it had a more permissive license.

This is the substantial argument. Before we see if Matt’s response pertains to this argument (Gruber claims it does not), it is worth looking, once again, at what Jalkut offers to demonstrate the greater success of a BSD style license.

Jalkut provides four examples (FreeBSD which Mac OS X borrows from, WebKit upon which Safari and Chrome are based, the version control system Subversion, and libcurl) of which one (libcurl) I will ignore in the interest of your time. For contrast, while acknowledging the massive success of WordPress, Jalkut lists in the GPL roster Subversion competitors like Git, Bazaar and Mercurial.

He admits to ignoring such wild successes as MySQL, GNU/Linux and gcc, and raises once again the doubt expressed above: “but we can’t know whether their success is in spite of the restrictions their license places on participation by the broader developer community”. Well, if so, we cannot know the converse either, that the GPL “stifles participation”, at least not from the examples and reasoning offered by Jalkut. If the GPL “stifles participation” (Jalkut’s words), then, asks Matt, wherefrom the wildly successful plugin and theme WordPress community? This is the relevance of Matt’s point #2 (as quoted by Gruber) to Jalkut’s post.

Gruber might respond: well, all Matt does is say that the community flourishes, not in spite of, but because of the GPL. How does he prove it? But as Jalkut notes, it’s a difficult, perhaps impossible, matter to answer that question empirically. The examples offered do not settle the question, especially not in the “more liberal” license’s favour. In lieu of empirical data, Matt offers his reasoning in his post, on why the GPL might engender certain “core freedoms and principles” that help foster a “passionate community” (Jalkut’s terms).

3. Jalkut wasn’t arguing in favor of more restrictive licenses; he was arguing in favor of less restrictive ones: BSD/MIT/Apache style ones.

True. But what was the argument offered in favour of the “less restrictive” ones? A significant argument was the idea that such less restrictive licenses would grow the “community” by attracting closed-source development. Jalkut even offers an example: Versions, the beautiful Subversion client developed and sold by Sofa. This is not a very productive line of argument, since (as I wrote earlier), Sofa’s corporate activity doesn’t grow the community’s riches by much (perhaps not at all, in a deep sense)… it’s a dead end, a terminal node, in the tree of community code.

Which is the point (#3) that Matt makes from his experience of watching the WordPress community go “dark”.

In other words, it is not convincing to bring up commercial end products, however well designed they may be, as examples of growth of community.

Incidentally, it is not at all clear to me that it is infeasible to build a Versions client (with the same sparkling UI, and with roughly the same effort) for a GPLed Subversion. That is part of the relevance of Matt’s point #1: WordPress’ GPL license has not jeopardised its acceptance by clients (corporate entities, I would presume) nor has it, obviously, led to a low rate of adoption in comparison to “more liberally” licensed alternative blogging platforms.

Conclusion

Jalkut’s post is motivated by Matt’s take on what the GPL implies for WordPress themes. Here is Jalkut’s quote of Matt:

One sentence summary: PHP in WordPress themes must be GPL, artwork and CSS may be but are not required.

From simple commonsense notions of fairness (which if I understand him correctly, Jalkut calls upon in his own arguments), this is eminently sensible and fair. PHP code in themes relies on WordPress capabilities, leverages the functions and mechanisms it (the WordPress API) offers, and adds not much significant original value (you loop through the posts, or comments, and you print them out, along with rendering some additional stuff like sidebars by invoking relevant API calls). On the other hand, your original contribution, the art and design implemented in the CSS or any artwork, is left to your whims to dispense under a license of your preference. Contrary to what Jalkut writes, a user of a theme cannot just zap any offending element and obtain the riches of the theme. Not without violating a license if so restricted by the author.

One other thing: It’s a hoary bit by now, that old half-quote: What is good for GM is good for America, but it’s counterpart seems to be in play in this debate. Jalkut, along with his even-handed prose and detailed explication, is also trying to “scare the bejeezus out of” you (by implying) that what is bad for Apple or Sofa is also bad for you. The GNU foundation’s argument is quite the opposite. And not much of Jalkut’s argument makes his case, in my view.

Postscript

I am an avid and mostly appreciative reader of Daring Fireball.