Blogs

RedMonk

Skip to content

What is Open Source? The Q&A

As justification for entries goes, “well, everyone else was doing it” probably isn’t going to win me any prizes, but that’s about the best I can do here. I’ve been assiduously avoiding the subject thus far, but I guess it’s time to jump off this particular bridge. Because, you know, everyone else is doing it.

For the record, I did make an honest effort to sit this one out, but I knew, more or less as soon as I saw it, that I’d eventually have to comment on Nat’s thoughtful dissection of the term “open source.” It’s sort of tough to be an “open source” analyst and not weigh in on the subject, much as I don’t like the idea.

So here are my thoughts and comments (such as they are) on the currently raging (so to speak) “open source” debate.

Q: First, any disclaimers to make?
A: Of the mentioned vendors, EnterpriseDB, IBM, MySQL and Sun are RedMonk customers, while HP, Oracle, and SugarCRM are not.

Q: Ok, for those that haven’t been following the discussion all that closely, can you bring us up to speed?
A: Well, a couple of folks have already done that. Stephe’s got an excellent roll up of the action so far here:

It all started when a terrible article (“Ten Leading Open Source Innovators“) prompted Nat Torkington to ask “Is ‘Open Source’ Now Completely Meaningless?” (The article also provoked Chris DiBona’s ire). Enterprise Open Source Magazine jumped into the fray with an odd article grabbing sound bites from various people and going under the sensational headline “How Open Is ‘Open’? – Industry Luminaries Join the Debate“. Michael Tiemann chimed in with an articulate response, Nat finished up a roll-up post, and it even provoked r0ml to blog again — always a good thing.

Allison Randall chimed in her contribution discussing EnterpriseDB. This provoked Matt Asay to fire back (fresh from Alfresco’s fine second relicensing move to the GPL).

The only thing I’d add to that recap is that I actually believe this debate has much deeper roots. To me, this is in many respects – as mjw asserted in #redmonk today – just a replay of the old “BSD v GPL: what is freedom?” debate with a new cast of characters. The issues here are strongly held, and resistant to compromise. Regrettably so.

Q: Why would you have skipped this discussion? What about it is unappealing?
A: A couple of reasons. First, it’s the type of philosophical discussion I try not to engage in, because it’s generally not possible to persuade people with fixed positions and because there really isn’t a right answer. And that’s actually number two: most of the people I’ve seen weigh in here are right. Nat’s right, Stephe’s right, r0ml’s right, Chris is right, and so on. It seems as if there are people here looking for a black and white, this is open source vs this is not, and I personally don’t see that as a realistic expectation. But lastly, I wanted to avoid it because outside of the immediate open source community, I don’t think anyone’s going to care. We can debate these issues all we like, but customers cannot be expected to parse the multiple levels of nuance and complexity that attend to these discussions. I could not agree with Stephe more that this is the very definition of a tempest in a teapot.

Q: So you don’t think Nat’s original entry asks a legitimate question?
A: It absolutely does, because it’s asked within the specific context of OSCON. As the man behind one of the most popular and relevant open source conferences, Nat absolutely has the right – if not the obligation – to explore the question on behalf of attendees, sponsors, presenters and so on. The ensuing debate, however, has broken free of that particular paddock to range fairly far afield, and that to me is a problem. When the debate is stretched to the software industry more generally, it becomes diffuse to the point of being irrelevant. In my opinion, of course.

Q: But isn’t the definition of what is and what is not open source important?
A: Yes, but I don’t think it’s particularly manageable. The attribution debate, if nothing else, seems to have made that point. While many might agree with Chris in criticizing the SugarCRM licensing approach, last I checked they’re still billing themselves – unapologetically – as an open source firm. They’re careful to use the “commercial” qualifier, but I don’t see them dropping the term, nor an external party finding sufficient grounds to compel them to do so. At some level, then, the debate is academic, not to mention not terribly interesting to customers.

Q: Accepting for the sake of argument that customers won’t care about these issues, do you think that’s smart? Should they care?
A: To the extent that it impacts what they can and cannot do with their software, yes, I think they should. But frankly, customers have enough trouble with even basic open source principles: asking them to digest the implications of attribution licenses, as an example, I think is too much to ask at the current time.

Q: Where do you, personally, stand on these issues?
A: At heart, I’m a pragmatist. I’m in favor of a larger and more vibrant open source ecosystem, so anything that furthers those goals is generally something I’m in favor of. While I understand and appreciate the rationale behind Chris and Nat’s frustrations, I don’t necessarily feel the same way. Yes, there are risks as Ian discusses, but I think they’re vastly outweighed by the potential contributions.

As an example, only a few years ago, I used to hear people lamenting the participation and involvement of large vendors like HP, IBM, Oracle and Sun within open source communities. I found this position incomprehensible, as it seemed self-evident to me that the resulting influx of resources, cash and attention was generally a positive. Not a universal positive, perhaps, but generally a benefit to the affected communities.

Q: How about what Nat terms a “tricky boundary case”, EnterpriseDB?
A: EnterpriseDB to me is more Rorschach than edge case; most will look at them and see what they want to see. If you’re of a more purist mindset, you’ll probably highlight the fact that EnterpriseDB merely builds on open source, maintaining their own non-open source codebase. Unlike, say, a MySQL. If you’re of a more pragmatic bent, you might point instead to the fact that the firm invests heavily back into the Postgres foundation they’re built on – work that benefits that community directly, and is open source. I could build a case for either side, but given my admitted pragmatic leanings, I’m personally a believer that EnterpriseDB is both an open source company and deserving of a spot at OSCON.

Q: But where do you draw the line; is Windows, as an example, an open source product because it incorporates BSD code?
A: No, it’s not. r0ml’s right to bring them up as an example, because it highlights how complex the questions can be when you factor in the BSD brand of freedom. But while I’m no more inclined to argue with r0ml than I am with Mark, I think this is to me is an obvious extreme. Even the most lenient pragmatists such as myself would not argue that any product incorporating open source code is an open source product. What’s more, I’m fairly confident that Microsoft would have little interest in branding it as such, so again the point is academic.

Q: How then would you qualify what is or is not an open source firm?
A: As can be seen in my distaste for the seemingly popular discourse on Web 2.0 definitions, I’m not a huge fan of strict, definitional discourse. I’d rather make the decision myself on a case by case basis; by eyeballing the firms, as they were. Much as I do now with Web 2.0 properties. This, of course, is an approach that is laughably non-scalable, so if pressed I’d simply retreat to my original contention: are the firms sustaining and growing the open source projects they’re associated with? Are they growing the open source ecosystem as a whole? If the answer is yes, then I’d be inclined to allow them to use the open source moniker with little or no restriction, because their success is the community’s success.

Now it’s true that if you can’t drive trucks through the holes in the above definition, you probably can’t drive, but I’m relatively unconcerned about that. Some might point out that IBM, under the terms above, qualifies as an “open source” firm, despite the fact that the vast majority of their portfolio is in fact closed source. Which would be very legitimate pushback, except for the fact that a.) IBM’s not going to brand themselves that way, and b.) if they were to use the term selectively I think it would be more or less appropriate given IBM’s contributions to Apache, Eclipse, Mozilla and so on.

Q: What about other dimensions to assess open vs closed?
A: Josh (a Postgres commiter) has done a good job of listing some of the facets firms or products could be measured in. For the purposes of the folks in the open source world, these are useful, in my view, for framing the debate. As I told the folks on #redmonk, however, I’d be careful in expecting customers or buyers to appreciate them. As Stephe says, most customers are concerned with one thing: solving their business problem. If open source is a better way to solve that problem, then great, they’re all for it. But pursuing the philosophy of what is and is not open source is not likely to be an exercise that holds a lot of appeal for them.

Q: What about SugarCRM? Many seem to feel that they’re another complicated edge case that’s unfairly blurring the lines of what open source means – do you agree?
A: The two principal complaints that I hear about Sugar – that they’re released under an attribution license of questionable OSI certification, and the fact that only a percentage of their product is in fact open source (I’ve heard everything from 60% to 80%) – are both legitimate. There are good reasons to dislike and/or disapprove of both decisions (although I do find it ironic that some of the same open source champions that cite Sugar when they need examples of application success are now criticizing their approach); I’m personally not a fan of either one.

But I do support their right to make those choices, and I do believe that customers are benefiting from the code that is available. The same is true of Zimbra; I’ve gotten pushback from a couple of people concerning Zimbra’s some features are free and open source/some are not strategy, and I quite understand it. If you’re looking for Blackberry connectivity, as an example, it is no doubt frustrating to learn that those features are not available in the free and open source product set. But I wouldn’t try and argue that it’s not a good thing that some of the features are available for free, and available as open source. I’d rather have 80% of something than 100% of nothing, in other words.

Q: Basically, it seems as if your approach can best be summed up as – don’t worry about it, even mostly open source firms are a good thing. Doesn’t that raise longer term problems, if more open source companies embrace approaches that don’t release actual code?
A: Only if you assume that a hybrid approach is inherently superior. Matt, for example, is concerned about the prospect of EnterpriseDB locking in their customers, but is at the same time quite outspoken on the inherent advantages of a 100% open approach.

I happen to believe that those would-be disruptors relying on the so-called hybrid approach are, to some extent, leaving themselves open to disruption. All things being equal, most customers when presented with a choice between Product A which is 100% open source and Product B which is 80% are going to choose the former. So over the long term, I tend to believe that that’s the direction that more firms will move in – Matt’s Alfresco being proof of this. It’s difficult, then, for me to get worked up about firms that want to take a halfway approach in the meantime, as long as they’re a net positive to the open source ecosystem as a whole.

Categories: Open Source.