“One country . . . one ideology, one system is not sufficient. It is helpful to have a variety of different approaches . . . We can then make a joint effort to solve the problems of the whole of humankind.” Dalai Lama
Open source and Linux is currently at the forefront of the new OSI Licensing Process and the GPL version 3. This usually would boil down to nothing for me and most likely boils down to nothing for most end users. However, in the cases of both of these ‘improvements,’ there are alternative motives. Don’t fool yourself into believing that all intentions of the OSI is to make Joe Common and his laptop full of Linux happy. While they may have Joe’s best interests in mind…they have his pot on the backburner while the new Teflon coated enterprise pot is heating up nicely on the front one.
You’ve heard me talk about this before in a couple of articles [1] [2] that attempt (albeit, my first attempt so please be kind 🙂 ) to address the issues that come into play when businesses and corporations assert their influence into open source communities and projects. Once again, I feel that not enough people are looking at the whole picture. Not to say that I have any ability to understand anything better than others; just that I make a consorted effort to always look at things in 3-5 different ways (leftover habit from college philosophy class).
What do I have a problem with this time? Let’s start with the OSI wanting to ‘trim the fat’ and drop or de-emphasize licenses. Why do we need this? Intel seems to think that it is needed and the community seems to think that there are too many licenses also. Why too many? Who says that this is needed? Can fewer licenses even be feasible?
I don’t think the OSI is actually thinking about things…they’re listening too hard to businesses and enterprise to be thinking. They’re listening to these enterprises and large businesses whine about having to hire X number of lawyers to sift through licenses to make sure that they know what they have to do in order to use the code. Notice I put emphasis on what “they” have to do i.e. the business or corporation. My reasoning lies in the fact that, by nature, a business wants to do as little work as possible to A) save money while at the same time they want to B) save time. In this case, they want to have their cake and eat it too. They want the benefit of few licenses to sift through and they want the OSI to do it for them…and for free. What’s the best way to do this? Have a keynote speaker (HP’s Martin Fink) from a company that is Linux-friendly (HP) give a call-to-arms at Linuxworld.
I don’t have a problem with businesses wanting to make more money or using open source to make money. I do have a problem though when businesses take open source for granted. They act as though Open Source was MADE FOR THE ENTERPRISE. We know this isn’t true. Open source was made for the common good of everyone, not just those in I.T. nor in the enterprise environment. In this case, I feel that the call to remove licenses from the OSI model is just that…taking OSS for granted. These businesses are assuming that they can have the OSI sort through the licensing schemes and ‘clean out the closet’ in a sense. And the OSI is letting them do it by actually listening to what these businesses and enterprises are saying.
So is this spring cleaning needed? Nope. It’s only recommended. Notice the people that are participating in said recommendations? Is it Jon Smith who uses OSS to develop software on his desktop? Nay, it is Intel and HP. It’s large corporations. It’s only needed by businesses that want to save money and get something done for free at someone else’s expense. The common sourceforge project doesn’t give two squats about licenses that they don’t use…and that is what open source is about. Not squabbling about a license model, but developing projects and making progress on said projects.
The beauty of the OSS model we currently have is that it has many licenses to choose from.
This allows software developers to choose what is right for them and to open source their software with a license that they are comfortable with and that their customers are comfortable with. Projects that might not open source their code will take a look at the MPL and say, “Oh, look, this fits our business plan” and they’ll go for it. They might not be willing to take the GPL plunge just yet. Getting rid of licenses means getting rid of choices and ostracizing developers and projects by NOT allowing them the choice of license…in a sense, deciding for them what they are allowed to open source and under what license they are to do it under.
Let’s take a look at an actual case where the above paragraph is illustrated. We might use Sun Microsystems as an example. Sun took Solaris and contributed the CDDL to the OSI license listing. While I don’t fully agree with the license Sun released this under (believing that the GPL would be better), I do agree with their intentions. They could be hoarding code like Microsoft, but they’re not. And they released their code under an OSI approved license…which is a heckuva lot better than nothing at all. That’s the beauty of having many licenses. Had they kept things closed, would open source have as much attention as it does now? This is only a small example.
Variety, they say, is the spice of life. What does the end user have to gain from the OSI license slashing? NOTHING. A big, fat zilch for benefits. Big business and enterprise would benefit from this because they wouldn’t have to pay Lawyer, Lawyer2, and Brownstone 200 bucks an hour to ruffle through the different licenses. Developers might think that it will be easier to sift through the licenses also. But in the end, NOTHING is gained by getting rid of licenses. We lose more than we gain…we lose choices and alternatives. I don’t know about everyone else, but when you lose more than you gain at something…you don’t do it. Perhaps the OSI should take a few steps back and reconsider which ear they’re listening out of.
Just a nitpick – the title of your article should be “Are TOO Many Licenses a Bad Thing?” The misspelling looks decidedly unprofessional, especially in a title that is picked up by other news services. 😉
With regard to the content of your article, those in favor of limiting the selection of open-source licences do so for nearly the same reasons as those who want to make a basic Gnu/Linux installation simple, without ten different desktop environments, three office suites, five browsers, and so on. They don’t want companies/people to shy away from using open-source (or Linux) merely because of the potential complexity of use.
While I personally prefer a multitude of choices, optimistically believing that the best will win out and weaker competitors will wither, I have to recognize that sometimes it’s publicity that creates momentum and not necessarily quality. The more choices, the more complex the decision, and the more likely someone will use something they’ve heard someone else recommend (or advertise) rather than evaluating each choice on its merits.
Therefore I cannot agree with your assertion limiting license choices equals a loss for the end-user or the community, and a win for the corps. I don’t think anyone can tell, yet, how it will play out – I think the potential for a win exists either way.
Thanks for the spelling tip and reply. I just hope they do things for the right reasons and not the wrong ones. :/
It absolutely helps the corporation, but it also helps the average hacker as well.
Free Software was not made exclusively for non-enterprise users. In fact, commercial interests were encouraged to use and create Free Software.
That said, you’re right, only a single open source license. or even only two, would do more harm than good.
But having hundreds of choices that are only different in minor ways isn’t great either.
The Intel Open Source License http://www.opensource.org/licenses/intel-open-source-license.php
and the NAUMEN Public License
http://www.opensource.org/licenses/naumen.php
are a case in point. Who needs both of these since they say the same thing? The only difference is whose trademark they are asking you not to use.
I don’t believe the average user will feel deprived of these two licenses, especially if they can be combined into a general purpose license.
English Nut makes a point. The free market isn’t necessarily free, so something of poorer quality can win out over something of higher quality simply because people don’t like change or never hear about the better license.
And in this case, lowering legal costs to develop and work with FOSS means that more developers, corporate or otherwise, will be using FOSS. If someone chooses the BSD license as opposed to the MIT license or the Apache license, are we really losing out?
If a company creates its own license to avoid the hassle of finding an appropriate license, we now have a new license to learn about. Is that really a greater gain than the loss in understanding and clarity from other developers? What’s the point of having OSI-approved licenses if everyone should just create their own for the benefit of having more choices?
The gain we have is the fact that other companies behind said company might open source their stuff under the same license…or they might do their own. By eliminating licenses, we’ll short people the ability to choose those licenses and we’ll discourage the creation of new ones….a “just pick one of the ones we have” mentality. If that were the case with software, everyone would run microsoft.
“By eliminating licenses, we’ll short people the ability to choose those licenses and we’ll discourage the creation of new ones”
You’re arguing that we need more licenses to encourage the creation of more licenses?
There needs to be a balance between a free-for-all, hack-out-your-own license mentality and a one-license-to-rule-them-all mentality. More licenses for the sake of having choice isn’t good enough, but neither is restricting everything to one license for the sake of ease. Both extremes have costs.