In an earlier post I gave my take on why Matt Mullenweg is spot on in his criticism of Dan Jalkut’s criticism of the GPL. Shortly after Matt’s post, John Gruber of Daring Fireball chimed in with his thoughts. First he quotes Matt, which I reproduce in entirety below (to be clear, this is Matt’s text):
- 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.
- 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.
- 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”:
- Jalkut wasn’t arguing about whether users will not use GPL software; his argument was about developers.
- 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.
- 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:
Some Basics
It’s worthwhile to pause and make explicit the underlying difference between GPL and other (“more liberal”) schemes. GPL is explicitly (and exquisitely) crafted to enable an alternative mode of development to the closed, corporate one. The motivation is not merely to grow the number of developers or the lines of code, but to create and sustain a community involved in such a mode of development. It is critical therefore to understand the difference between such a philosophy and the one espoused/defended by Jalkut, in passages such as these:
This so-called “viral nature” of GPL is what scares the bejeezus out of companies, large and small, who fear the consequences of having to give up their own intellectual property to the public.
[emphasis mine]
It must be mentioned that Jalkut is a fair chap, but in attempting to represent his opponents thus:
I know what some of the GPL-enthusiasts are thinking now: “leeches don’t count as community.”
he might as well leave the name-calling (“leeches”) out and address the content/question: what counts as community? Instead, he skips right ahead to the question of how to get “people in the door”. The contra-Jalkut might respond: “corporate coders do not count as people”!
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.
Leave a Reply to Josh Betz Cancel reply