During the first years of Java's phenomenal growth in popularity, I fully endorsed Sun's controlling authority over Java. Sure, Sun did it in part to make money on licensing (as if that's a bad thing), but the bottom line was that Sun had no choice but to assert its control.
Microsoft was itching to distort Java into a Windows-centric language, thereby undermining Sun's effort to promote Java as platform independent. Regardless of whether you think Sun was successful at making Java platform independent, the only way Sun had a prayer of achieving its goal was to prevent companies such as Microsoft from coopting Java.
Had Java already been widely adopted, Sun could have relied upon the fact that a platform-independent Java was the de facto standard. But it hadn't been widely adopted yet, so it didn't have the status of a de facto standard to protect it. Sun could have turned Java over to a standards group, but it could not do so without risking the possibility that Microsoft would influence that group in its favor. Considering Microsoft's bank account, that was a very realistic fear. Therefore, Sun's only viable option was to retain control over Java.
Things have changed, especially in the past year, and Sun is recognizing that it can now operate with less fear of Microsoft's influence over Java. As a result, Sun is now considering releasing Java under the Sun Industry Standards Source License (SISSL, pronounced "sizzle"). I'm not sure anyone fully understands the ramifications of the SISSL yet, but most people speculate that it will allow Sun to have final say over what constitutes the official version of Java.
Before I continue, let me make it clear that I believe Sun has the right to assert any level of control that it desires over Java. I also believe that Sun has the right to charge what the market will bear for Java licenses. And if Sun decides to continue that route, I would defend its right to do so.
But for strategic reasons, I strongly recommend otherwise. Instead, I would like to propose that Sun go for broke and license Java under the GNU General Public License.
What are the drawbacks of doing that for Sun? Depending on how SISSL works, the GPL might reduce or eliminate Sun's ability to make money selling Java licenses. But that conclusion is premature, no matter what SISSL amounts to. Troll Tech has created a triple license model for its programming toolkit Qt, including GPL as one alternative. Troll expects to continue to make money on its commercial licenses, regardless of its GPL alternative. Whether that is possible remains to be seen, but Sun could try a dual or triple license as well.
The advantages to making Java available under the GPL are much easier to enumerate. First and foremost, a GPL Java undermines the supremacy of Windows better than Java under any other license. It also disrupts Microsoft's model for making money on its latest Java competitor, C#. There are still a number of reasons people might want to choose C# over Java but, as Microsoft demonstrated with Internet Explorer, very few things can stand up for long against free competition -- especially when the free product shows up on every copy of the operating system whether you want it or not.
Granted, Sun's Java is likely to ship with all Linux distributions by default, even if SISSL is simply compatible with GPL. But Sun can initiate much more momentum for Java on Linux -- the platform in which Java probably has the least momentum right now -- by using the GPL itself. Sun needs the open source development community, not the user or administrator communities, to accept Java. And like it or not, the open source development community clearly favors products that are under the GPL.
More importantly, however, the GPL prevents companies such as Microsoft from modifying the source code without redistributing its modifications back to the community. That is anathema to Microsoft. Unless there is a change of guard at Microsoft, licensing Java under the GPL would be as good as issuing a guarantee that Microsoft will never touch Java again.
A Motif for the future
To most Linux developers and users, Motif is essentially irrelevant. But that is only because the Linux community tends to focus on new software for Linux. And when it comes to writing new software, the most popular choices for graphical toolkits are GTK and Qt.
One thing the Linux community needs to understand is that continued Linux momentum depends at least partly upon the willingness and ability of commercial vendors to port their Unix products to Linux. It's all well and good to shout down the commercial database competition with claims that MySQL or Postgresql is or will be competitive. But Linux advocates would be ill advised to forget that the endorsement of companies such as Computer Associates, Informix, and Oracle gave Linux the credibility it needed to get out of the closet and into the enterprise.
The point is that companies, including Informix and others, often depend on Motif's presence on a Unix system in order to sell their wares. Lesstif, the free version of Motif, is insufficient for their needs.
That topic came up at an LSB meeting, and the participants widely agreed that LSB could not make Motif, or even Open Motif, part of the LSB standard due to licensing issues. As a side note, I find that to be a needlessly artificial restriction. If LSB wants a GPL-compatible standard as a base, nothing prevents LSB from defining multiple supersets of that standard, one or more of which could include software that is not compatible with the GNU General Public License.
Regardless, in reality LSB is unlikely to include things such as Open Motif in some future standard unless the license becomes compatible with GPL. So consider that a petition to license Open Motif under the GPL.
I'd like to direct that request to The Open Group, which is the body that initiated the change to the Motif license in the first place. But I'm almost certain the process is much more complicated than that.
I am not a lawyer, and I do my best to stay away from lawyers unless absolutely necessary because I'm afraid I might get a brain hemorrhage trying to figure out what they're saying. So I'd rather not try to unravel exactly what process The Open Group has to go through to change the Motif license to GPL. I assume such a process would not only be difficult but lengthy, since The Open Group is not the only entity involved. But I encourage The Open Group to take that direction immediately if it hasn't already.
Final thoughts
The one thing I do not want my readers to take away from this column is that I favor licensing all software under the GNU General Public License. I believe the GPL is a good thing in many cases, particularly when you're talking about the kind of software that can otherwise be used to restrict developers' and users' choices. In those cases, I enthusiastically advocate using the GPL. It is also admirable when developers want to ensure everyone can use or modify their work -- provided the developers can afford to do so.
But those are not the reasons I advocate using the GPL for Java and Motif, as you'll see if you examine my arguments above. In the case of Java, I see the GPL as primarily a strategic weapon to advance an excellent platform. In the case of Motif, I advocate using the GPL as a strategy to get more commercial software developers to support Linux.
I am not attempting to make a philosophical statement about the GPL. Use it when it makes sense to you, whether your thinking is philosophical, strategic, or philanthropic. And I urge you to use any other license you wish when you think it makes sense -- including BSD-like licenses and proprietary commercial licenses. Let the market decide if you've made the right choice.
