Skip to content

The Open Sourcer's Dilemma, or, The Long Post on The IBM Open Source Analyst Briefing

As mentioned previously, I was at the IBM Open Source Analyst Briefing earlier this week. My overall impression is that IBM is doing the right thing when it comes to working with open source. My only qualification to that — and it’s a huge one — is that IBM still has to make money, and lots of it. This means two things: (1.) IBM is never going to please all of the open source people out there, and, (2.) there’s still a lot of eggs in IBM’s closed source software basket.

The Right Thing

What is “the right thing”?

  • Being as transparent as possible,
  • Allowing employees to work full-time and part-time on open source,
  • Open sourcing code (that’s more than just “Waste Dumping”),
  • Using open source software (while OSS Waste Dumping rings true with me, I prefer to call the concept of “OSS Strip Mining” by another word: “programming”),
  • Maintaining a conversation and doing work for with people who are interested in open source.

As you, dear readers, probably know, I’m not much for discussing how many angels can fit on a the head of a pin. The debate about open sourcing Java, for example, isn’t that interesting to me past the initial stalemates of “it would be better!” and “no it’s fine!”. So, I’m sure there’s endless discussions we could have about who’s church of open source is the best and which people and organizations attend the right church and sing the best.

Super. Enjoy your early Sunday morning hymnals. I’ll be sleeping in.

The Open Sourcer’s Dilemma

While it would be valuable to write-up my notes realtime style (as I’m fond of doing), I’d rather spend time cracking the nut that I keep coming back to when I think of open source: does source code alone have any monetary value? If it does, then selling it “works” (as it currently does) if source code doesn’t have value on it’s own, other models are needed, and soon.

The basis of those model is having both the vendor and the customer realize that “free” software can’t be created for free. The dilemma is that we (the software industry) might have set the bar way, way too high for how much can charge, and we’re going to suffer when that bar falls.

The Billion Dollar Build

I spent a long time on the plane last night trying to come up with a sort of “calling bullshit,” cynical angle on the event, but in the end I felt like I was just trying to look smart instead of right. As I said above, IBM looks healthy in the OSS world.

That means there isn’t much to say about IBM and OSS: you’ll hear plenty from them, so there’s no need to duplicate it here. But there’s a lot to say about the software industry and OSS. Here’s the only salvageable material along those lines I ended up with last night, jammed in here gonzo-style:

No matter how you slice it, here’s the big question: how do you put value on actual software itself? The first copy of any piece of software can cost millions of dollars (arguably, billions if you roll-up historic costs with something like Oracle or Windows). I won’t bore you with a list of how that money is spent.

The majority of you, dear readers, along with me, are the major part of where that money goes: we have the gadgets, houses, cars, families, and nice lunches to prove it. Yay capitalism! It’s spent on you. (See name of this blog ;>)

Not so much on that second copy. The second copy costs nothing.

And there’s all the trouble. And the money.

Without charging for that second, third, and n-th copy, none of us would be here reading this blog or typing it on a jetBlue flight watching some ridiculous advermentry for Aquamarine. Like the screens on this plane say, “without you, we’d just be flying a bunch of TVs around the country.”

The point is: I would be nuts if I said charging for the second copy was crazy. And I’d be stupid if I thought it wasn’t a valid thing to do given the context we currently live in. A “truth” of software as it were…albeit on that can change.

Software wants to be free, sure, but I’m sure it wants to run on a nice PowerBook too. If I know anything about PowerBooks, it’s that they don’t want to be free.

Yes, friends, we need to make money. As James was saying this morning, baby gotta eat. TCB.

(For some great thinking out loud on selling software, see Niel Robertson’s “The Increasing Tail”.)

Support vs. Extort

In my mind, the discussion surrounding Creative Commons is a good analog to the closed vs. open source issues with software. Creative Commons doesn’t start out with the notion that people should be able to freely copy things, no questions asked. It accepts the idea that people want to get paid for things they create. This is in stark contract (real or incorrectly perceived) to the GPL.

What CC doesn’t accept is a model that only encourages people to extort that money forever (see the never-ending copyright on The Mouse). The belief is that, eventually, anything that’s created belongs to the group and not the individual. I might be mixing pure Lessig-think with canonical Creative Commons: frappé!.

The creators can choose how to support their lives, but they can’t and shouldn’t want to get a strangle-hold on the group due to copyright law. We need to support the creators and in return the creators will take a pass on extorting us.

Closing in on Closed Source

To come out and say it: I don’t see much long term future in doing exclusively closed source approaches to software. The reason is as simple as “the cat’s out of the bag.” When you have companies open sourcing applications as traditionally closed source as CRM and PBX, it’s game over for the last, best hope commercial software had: The Lefkowitz Theory of Cool Projects Only. (With no discredit at all to r0ml: that was damn insightful thinking.) Closed source becomes a weakness because another company can take an open source approach and cause endless trouble for the incumbents.

The problem is that software in itself might be way over-valued. That’s the threat that JBoss brings. Their revenue is high from an individual’s perspective: $20-50 million is one estimate. But, that’s nothing compared to IBM’s, Oracle’s, SAP’s, Microsoft’s, or any large closed source shop’s revenue. If JBoss, or whoever else, can sustain itself on sub $100 million revenues and keep up the pace of features and quality, multi-billion dollar companies that rely on software and services will only have time to save them: the time it takes those small, OSS companies to catch up technologically.

(Of course, the chance of those small companies getting greedy and burning all their cash is high: no one ever learns that lesson, no matter how many bubbles we live through.)

People are Affordable, High Margins Aren’t

The real value in software is the people involved with it: those who create it, support it, and run it. That’s the secret of OSS companies: if you satisfy the people — which doesn’t take much compared to what it takes to satisfy The Market — they’ll pour their hearts into writing software.

So, if you’re a JBoss, you can make much less than an IBM if you redirect most of the money back to your people. That approach can get the same technological results, meaning you can charge customers much less for the same thing. See Linux, Apache HTTPD, and most recently, the IDE market. That’s the deep fear that open source elicits from closed source companies: OSS can do it for less, much, much less. IT won’t go away, it just might get really cheap because the groups creating it might need just $5 million instead of $5 billion.

End-user Driven OSS

Least we get all wrapped in up fretting about vendors and coders, another approach in an OSS world, which we haven’t seen much of, is customers getting much more involved in patronage and pooling their money together to develop and support software. Sabre comes to mind, but to be completely honest, I don’t know enough about it to say it’s a good analog, even ball-park wise.

There’s a huge vacuum around end-user involvement in creating software that is just begining to be understood, let alone addressed, thanks to things like Greasemonkey, Ning (which came up a lot in Steve and I’s talks with IBM’ers), MySpace, SalesForce, and other sites that offer platforms or tools that can customize existing systems and create new ones. Those are nothing to bet the farm on (yet). End-users “writing code” is a perennial favorite pie-in-the-sky, for sure. But those technologies and others are new and helpful ways of approaching the problem at hand: how does software and the people involved in it prosper in an OSS world?

Back to IBM

No one can really say what will happen because the two models — closed and open source — are so dramatically different. You can bet that commercial companies will fight tooth and nail to maintain the same revenue and growth: their hope is to figure out other things to charge for, as many of them are doing.

What’s important is proving your metal when it comes to navigating those waters without sinking. IBM’s messaging on that front is fantastic in it’s honesty. Person after person stood up and said, literally, we have no idea where this is going in the next 10 or 5 years. But, they have an advantage over many other people: they’ve accepted open source, they’ve got good models and thinking, and real, existent experience in open source. Oh yeah, and they have plenty of money and revenue to buy time to transform.

None of that is a guarantee, but it gives them a much better shot than others. It’d be a good sign if we keep seeing sentences like “we have done much in recent years to transform our company” coming from IBM.

Of course, there are plenty of other companies that seem well poised. I’ll have to see if their kool-aid tastes as good ;>

Disclaimer: IBM and Microsoft are clients. Others mentioned or linked to are not.

Categories: Enterprise Software, Open Source, Programming, The New Thing.

Comment Feed

2 Responses

  1. Good post. Coupla comments, briefly. First, you say what IBM does for open source, but not what open source does for IBM. As you so rightly say, its not about being altruistic. My understanding is in the open source game for a number of reasons:

    – to undermine Microsoft
    – to get pull-through into services
    – as on-ramps to more "enterprise class" offerings – DB2, Websphere
    – to look good in the community
    – because of the genuine benefits of the open source model.

    I'm not saying they're being bad, just canny!

    Second, one of the main drivers towards open source is the continued and increasing commoditisation of software. Linux is Unix, whether people like to agree or not, and that's been around for 35 years. LAMP is a commodity stack so its very hard to get any money out of it, the alternative is to open it up. That's a very short sumary of a very long discussion, but it'll have to do for now 🙂



  2. As you point out, OSS Waste Dumping is a subset, of sorts, to open sourcing code. I consider OSS Strip Mining to be a subset of “programming”, as you call it.

    Every FOSS community wants their technology to be used as much as possible. Taking a FOSS technology and using it within a larger solution is fine if the tech is kept intact. It becomes Strip Mining when the tech is “Blue Washed” or blatantly forked with the intent of never contributing back changes to the original FOSS community and/or never releasing changes as part of an OSS product. The FOSS community deserves to hear and see the thoughts and ideas for change, otherwise their tech splinters into many different permutations (ala UNIX) causing headaches, supportability issues, and more work in general for all involved.