Blogs

RedMonk

Skip to content

When is Open Open? And When is Open Closed?

Frankly, it strikes me a philosophical divide akin to the one that exists between the Apache camp, which favors permissive licensing terms, and the FSF, which advocates reciprocal-style alternatives. Not solely in the depth of the divide, in fact, but also in its persistence.

The poster child for the debate, typically, is MySQL. MySQL, you’ll recall, is a project which is an open source project (we can have that debate later) developed almost entirely by a single vendor. Unlike projects such as Linux, which are developed collaboratively by multiple entities, MySQL alone bears the burden of development, in return for which they retain rights that developers of Linux do not. Among them, the ability to relicense the code under more liberal terms than the standard GPL – the foundation of the so-called “dual license” model.

Personally, I believe MySQL is both an open source project and an open source company, but opinions obviously differ. This opinion was outlined in an earlier piece written about OpenSolaris. Responding to a question as to whether or not Sun was an “open source company,” I said the following:

To me, that question, and much of the discussion above, boils down to a philosophical question: are Linux and MySQL both open source? Or do you believe that anything short of the Linux model does not qualify? That was, in fact, the question I asked Michael [Dolan] in a comment. His reply was this:
I do not consider MySQL to be an open source development community which to its credit, MySQL has never claimed; unlike Sun/OpenSolaris.

While noting that the response doesn’t actually answer the question, it does point to the fundamental disconnect between our two viewpoints: I’m judging open source by the availability and licensing of the code, while Michael, Amanda and other critics seem to prefer a community based metric. As should be expected, since it favors their platform of choice.

For my part, I believe that MySQL is in fact an open source company and an open source project, in spite of the fact that the development of the codebase is not open but rather done entirely (or nearly so) by MySQL employees. So, assuming that Amanda is correct and that there are 70 external OpenSolaris engineers versus Linux’s 3K, that still leaves OpenSolaris as “more open” than MySQL. Whatever that means.

The question of whether or not some projects governance and contribution models are more open than others is not, to me, the question. Or even a question, to be honest: of course there is a spectrum of open, and a preponderance of different models servicing quote unquote open projects.

The questions, to me, are first whether or not this diversity of approaches is a good thing or a bad thing for open source, and second whether or not the distinction is important to the end users of said projects.

As to the first, my typical response is to point to the LAMP stack which collectively has enjoyed such success, but individually uses a variety of different licenses and contribution models. It includes both permissive (Apache and PHP/Python/Perl) and reciprocally (Linux, MySQL) licensed assets and it incorporates single (MySQL) and multi-entity (Apache, Linux, PHP/Python/Perl) developed open source projects, and in doing so highlights the benefits of an approach that transcends philosophical boundaries.

Are those boundaries important to end users? In certain cases, certainly. Following the acquisition of MySQL by Sun, as an example, many contemplated doomsday scenarios in which an evil Sun “closed” the MySQL source going forward. Given that MySQL employs a high percentage of the developers in the world that are qualified to work on the code, a closing of MySQL’s source – ignoring for a moment the fact that such a decision would immediately render worthless the billion dollars Sun invested in the firm – would indeed pose some problems. But because the source for the software is…wait for it…open, these problems would be recoverable.

Is the development model that Linux operates under more open than MySQL’s, then? Indeed. Does that make the code itself more “open?” It’s debatable. Even if it were, how would one measure the distinction?

My analyst colleague Matthew Aslett from the 451 Group suggests that the idea of measuring projects companies considered open by the OSI definition according to a second set of metrics, by which they could earn the additional appellation of “Equitable Open Source,” might go some way towards defusing the debates about what is and is not open source. I respectfully disagree.

Those in the industry that might care have, I would argue, already formed their opinions on whether or not a project such as MySQL’s is or is not open source. And those outside the industry, well, I don’t expect they’d care. At all. Most of the enterprises I speak with are still struggling with the basics of what open source is and what it means. Intending no disrespect, most are akin to the robotic arm that was told to place blocks one on top of another to build a two foot tower, and began by dropping them from a height of two feet because it was unfamiliar with the concept of gravity.

I for one would not look forward to the challenges of educating them on a spectrum of equitable vs (? inequitable?) open source. Or organic vs non-organic.

I’m not arguing that there are no differences between the approaches, please note; merely that I think an appreciation for them is utterly beyond most of the end users I know, who see few if any problems with MySQL-style models.

Consider, as an example, MySQL versus Postgres. Both are open source relational database projects, one is permissively licensed and the other reciprocal, and one is developed by a single commercial entity and the other by a community that includes multiple commercial entities. Certainly there are cases where one is selected over another because of one of the above factors, but far more frequently in my experience the choice is made according to technical criteria.

Which, in my opinion, is as it should be, as I believe that diversity in both licensing and governance models is a positive rather than something that should be stamped out.

But as always, your mileage may vary.

As an addendum, for those (few) that are interested in the above subject, it’s looking as if I and some other folks will be sitting on a panel on OSCON at Friday morning to discuss it. Can’t find an abstract as yet, but will add it as it becomes available.

Disclosure: Both MySQL and its acquirer Sun are RedMonk customers.

Update: Corrected the Equitable Open Source target from projects to companies, per mtg’s comment below.

Categories: Open Source.

  • http://blog.milkingthegnu.org mtg

    Hello Stephen,

    A few remarks:

    – Matthew’s post (and mine for that matter) are about labeling/evaluating companies, not open source “projects” per se.

    – That’s precisely because of the diversity you’re noting and because of the struggle of regular companies with open source you mentioned that more indicators have to be built so that to help deciphering the FOSS landscape.

    – “Equitable” is certainly something companies would look at for two main reasons: on the one hand it could be a measure of how well a company takes care of its community and a such an incentive (or a disincentive) for external contributions. On the other hand, fair and ethical trade is becoming mainstream enough that even marketing departments are taking a second look.

  • http://redmonk.com/sogrady sogrady

    @mtg: thanks for the feedback.

    1. appreciate the correction; it’s been noted in the piece. i do think the distinction is less than obvious in some cases – MySQL being one – but you were right.

    2. as stated above, it’s not that i don’t recognize the distinctions that you’re referring to, i just am skeptical of the idea that more labels will help decipher the landscape. there’s only so much appetite for distinctions and differentiation; beyond a certain point, it becomes a limiting rather than enhancing factor.

    3. while i agree that fair and ethical trade is indeed becoming more mainstream via things like Corporate Social Responsibility, i would disagree that it can be evenly or consistently applied in this situation. to keep beating the MySQL horse, what is less “equitable,” let alone ethical, about their model versus Linux from the perspective of an end user? i’m not sure the distinction would be appreciated by those consuming as opposed to developing open source, which i’m guessing is the primary audience.

  • http://www.michaeldolan.com Mike Dolan

    I think there is a distinction worth noting in my comments and perhaps it will help clarify the point I think you’re trying to make. First, if you look at who’s commenting in “the community” and the perspective/function of that person you can see there are some who approach the “debate” from various POVs. First there are a camp of strategists who are trying to drive broad industry impact, an open standard, a competitive challenge. Second, you have those who are interested in getting the most people and the best code into the project who are more concerned with questions like what license lets me drop in the most code or conversely what license keeps my code my own but kindof open. Third, you have people who approach the debate from the POV of product management looking at a) how do I get the most users, b) how do I get X% to pay me and c) how can I keep a development community interested. These are just a few of the camps you might come across and I’m sure there are others.

    None of them are wrong in their POV. I personally have never stated MySQL has taken the wrong approach, done anything wrong in its community, and personally I’ve always been a strong fan of MySQL’s approach. They have been extremely up front and blatant about their business model and it’s always been an easy “take it or leave it” approach. But, you’ll see different people in the community react differently. Strategists might view MySQL as a broken open standard. Technologists might disagree that the licenses closes MySQL off from all the Apache licensed works and the approach MySQL takes towards community members contributing back. Finally, the product management oriented people might criticize MySQL for not capitalizing on its huge user base.

    I think what the world needs is a framework for looking at this issue. I have yet to come across one that works.

    I would also add that while only a small minority may understand the intricacies of the debate, that small minority are often the thought leaders many look to for direction.

    Finally, I have been openly critical of the approach to some projects as have many others around the world. Part of the issue has nothing to do with how its community operates or how a vendor manages “the product”. The issues arise when there’s a disconnect between what the company says and what it does that can become an issue. When that disconnect occurs, then you get all the camps I’ve described in a flurry and it inevitably speculation spills over to other areas that are going along just fine.

    However, that disconnect is exactly what the broader group of users and companies want to know from the minority who do understand the details. I also believe that minority has a duty to explain the fine print clearly. If this wasn’t of value, then I suspect you, me and many others who visit here would have a tough time justifying our jobs ;-)

  • http://www.michaeldolan.com Mike Dolan

    I should have also stated that “open open” and “open closed” is not dependent on what anyone actually does (which license chosen, accept contributions back or not, etc). What matters is the stated intent, goals and the strategy of any corporate sponsor that is the sole driver of the community and whether the actions, the model, the “openness” of the IP, and the decisions made match the intent.

  • Pingback: Perl Coding School » Blog Archive » perl code [2008-05-19 18:38:25]

  • http://www.webmink.net Simon Phipps

    I think it does matter to the end-user, Stephen. One of the important aspects for a whole bunch of adopters is the avoidance of lock-in. It’s undoubtedly the case that many people select open source software in part to avoid lock-in, and it’s my belief that their confidence is misplaced in many cases since the only tool available to break lock-in is a full-scale fork which, though technically possible, is unworkable for end-users.

    The other dimension of this issue is the liberties of community developers – probably outside your parish, I realise. As I’ve stated elsewhere, I believe OSI-approved copyright licensing is no longer a sufficient metric for the protection of developer liberties. I believe we need an “OSD” for each of the other “pillars of IP” – patents, trademarks and trade secrets.

    One last pointer: Have you take a look at Certified Open? http://www.certifiedopen.com/

  • http://redmonk.com/sogrady sogrady

    @Mike Dolan: i have no major issues with the lens you’re using to view the issue, nor with the contention that someone – the minority, in this case – needs to understand and appreciate the issues involved. where we might break – i’m not positive – is that i think the issues are somewhat understood by the minority, they are just not agreed upon.

  • http://redmonk.com/sogrady sogrady

    @Simon Phipps: i agree that lock-in is an important advantage of free and open source software. i further agree that it’s probably unrealistic that an end user could successfully fork, say, MySQL and expect to be successful maintaining it.

    but i disagree with the implicit assertion that the end user is the only or even the logical choice for the role of fork maintainer. if MySQL, for example, closed the source for its database, i would expect a service and support organization to spring up around the codebase more or less overnight, obviating the need for end users of the code to maintain the fork. moreover, i would expect said organization to be reasonably to highly successful in recruiting disaffected MySQL defectors as they left an organization they no longer believed in.

    as for your other comment, i’m not quite sure why community developers needs and interests would be outside our purview, but i’m certainly not unilaterally in favor of mechanisms that restrict them (the developers or their liberties). that said, i understand that many corporations – like your own employer – benefit from just such items (e.g. JCA’s) and that they are often a necessary evil.

    as for the contention that we need additional definitions around patents, copyright and trade secrets, i’d welcome that – but have doubts as to the ability to digest all of the potential implications. if you ask many of the developers contributing GPL’d projects to Sourceforge or GPL’d plugins to the WordPress Codex, many of them select the GPL not because of its merits but rather by default. as such, i think it’s not entirely reasonable to expect an appetite for the digestion of the nuances of the other IP pillars.

    but if it can be done simply, say Creative Commons easy, i certainly would not complain.

  • http://www.webmink.net Simon Phipps

    Thanks for the reply.

    > if MySQL, for example, closed the source for its database, i would expect a service and support organization to spring up around the codebase more or less overnight, obviating the need for end users of the code to maintain the fork.

    Well yes, and I dare say it would be formed by a number of idealistic former employees of said company!

    But let’s say it’s not that bad. Let’s assume the company in question (I’d rather keep it generic if that’s OK – CIQ) decides to stay robed in the clothes of GPL conformance, but acts in a strongly proprietorial way towards customers, exploiting their locked-in-ness more and more as they get deeper and deep in their adoption. Or let’s say CIQ decides to restrict commits only to employees and partners, so that there’s not quite enough outrage or dissent to seed the fork. Is the customer disadvantaged? I believe so in both cases. Should analysts advising the customer be concerned? Again, I believe so.

    This black-and-white view of forking is in my view an illusion. If Oracle can’t successfully fork Red Hat (my view, naturally, your data may vary), what chance does a startup have of forking CIQ’s software given CIQ’s commercial success and market stranglehold?

    > i’d welcome that – but have doubts as to the ability to digest all of the potential implications.

    Yes, it will need working through. I believe developers will be well able to look for red flags in a health report, however.

    > i’m not quite sure why community developers needs and interests would be outside our purview

    Just am impression gained from your use of the end-user perspective when we’ve discussed this in person :-)

  • http://ianskerrett.wordpress.com Ian Skerrett

    I must be one of the small minority but I actually think you can predict a lot about an open source company and project but how ‘open’ they are. The degree of openness has an impact on how the company relates with end users but more importantly other companies/adopters that might build value-add services.

    For end users, I believe Simon is correct. For successful open source companies/projects it is very difficult to have a successful fork. Trademark and copyright laws provide a lot of protection for said company.

    However, imho the degree of openness is more important for other companies or adopters that might want to build value-add services and products on the technology. If the open source company is being very protective of trademark usage or of product plans, they won’t nurture a strong after-market of products and services. This in turn weakens the support network for end-users.

  • Pingback: OSCron

  • http://www.linkedin.com/profile?viewProfile=&key=3 James Dixon

    I have no problem with the idea of a system for rating the approach that commercial open source companies have towards their community. The hard part is doing it well so that it does not unduly favor any particular licensing model or one specific part of the company life-cycle. The life-cycle part is hard because commercial open source companies go through multiple phases: adoption-only, adoption/customer, pre-acquisition, post-acquisition. During each phase they might take a different approach to things.

    Another issue is the audience. There are multiple audiences for this information: contributing developers, other contributors, and users. The needs and concerns of these groups differ and each individual can fall into a different group on a per-project basis. I personally am a developer, but I am a user of some projects, a code contributor to other projects, and a contributor to yet more.

    The issue of audience is important because most commercial open source companies want to increase the amount of hardening (non-code) contributions they get but need to carefully manage code contributions to the core-IP. This means that your impression of how open the company is will depend on what you are trying to contribute.

    James

  • Pingback: 451 CAOS Theory » 451 CAOS Links - 2008.05.22

  • Pingback: 451 CAOS Theory » Customers versus users: a distinction

  • Pingback: 451 CAOS Theory » Judging open source business models