Alex Hudson

Thoughts on Technology, Product, & Strategy

Asay and Tiemann, mano a mano.

Matt Asay has written another entertaining blog piece on his particular theories of open source economics, and Red Hat’s Michael Tiemann and he have engaged in what is superficially a bit of “Is not!” “Is too!“. Looking a bit deeper, though, it’s not really the pragmatics vs. the Stallmanites, even though that’s how Asay frames it.

Fundamentally, Tiemann is right on the money: a simplistic “supply and demand” view of how prices are set in a market place completely ignores the value that Red Hat offers to its customers. “Subscription” versus “box price” is not simply a semantic difference – indeed, that’s essentially labelling their customers as brand tarts unwilling to risk CentOS / Scientific Linux, and reduces the business decision to a simple money figure. That’s not how business works; the difference between “cheapest” and “best value” is huge.

Asay also bizarrely labels Red Hat a “distant second to Canonical” in the purity stakes. This is Canonical with the proprietary server management, proprietary file sharing, proprietary application store, etc.? I don’t even vaguely understand the argument here: either Matt is badly misinformed, or is just being very selective – the only thing I can think that Red Hat withholds is permission for others to use its trade marks. Which Canonical also does.

Then comes the claim that “The bulk of the best, most widely used open source is funded by proprietary dollars.” – followed by a call of thanks for the likes of IBM, HP, Intel. No doubt those companies do contribute a reasonable amount, but to credit them with the bulk of the best: that’s really stretching it. If you look at the actual factual information of who contributes what to projects like Linux, corporate interest is large, but “funded by proprietary dollars” – haha. What Asay is basically implying is “proprietary sales are underwriting the development of open source” – presumably some kind of mass corporate hallucination that has turned these businesses into charities, and pragmatism be damned.

Of course, the reality is these businesses would never underwrite development of software which wouldn’t make the money back, and indeed IBM’s vaunted “$1 billion investment” was apparently recouped in a single year. According to Matt, we should be thanking IBM for doing this: to my mind, IBM should be thanking the community for the contribution that has enabled it to recoup its investment so quickly (since 2002 presumably it has been making good money, too).

What Matt doesn’t seem to get is that this split-personality marketing of “we do all this open stuff, except for this scarce bit we’re charging you for!” is a prize example of a house divided unto itself. You can’t sensibly talk about the benefits of open source without contradicting yourself completely when it comes to the paywall behind which your proprietary software sits: basically you have to fess up that the open source bits are the bait.

What Michael’s post illustrates nicely is not just a clarity of purpose, but a 100% commitment to what they tell their customers: no ifs, no buts, but a single compelling story. Customers understand the value they offer, and that’s why they make money.

[Edit 20:24: just for clarity, my comparison of Canonical to Red Hat is not to denigrate Canonical: merely to illustrate that claiming Red Hat are a ‘distant second’ to Canonical in the purity stakes is utter nonsense. Also, my reference in the comments to “proprietary application store” should be parsed as “a store that hosts proprietary applications”, not “an application store that is proprietary”]


WikiReader – “Project B”


RMS, KEI and ORG tell DGC “No” on ORCL MySQL


  1. Look at the facts. Every software company has some proprietary bits. I think Canonical is FAR more open than RedHat in that regard as it is only holding some server bits back in the products you mentioned for various reasons so I suggest you open your blinders to what is actually closed in the Redhat world and you will find a lot more than you want to find.

    Canonical is quite transparent when it comes to what is open and what cannot be opened and EVERYWHERE it can, it opens – look at Launchpad.

    Redhat is NOT 100% open as Michael has implied. You have to buy a license in the form of a subscription to use RHEL. So where’s the openness there?

    Nice to see Redhat getting a dose. And nice to see they feel compelled to defend – maybe they are feeling the pressure.

  2. Alex

    Sure, look at the facts. What does Red Hat hold back? The trade mark. You don’t need a license to use RHEL, it’s freely available – you just don’t get to call it RHEL. You don’t think Oracle went and bought a truck load of RHEL licenses to release Oracle Linux, do you?

    Not sure which of us actually has the blinders on at this point. RHEL is not proprietary in any sense of the word, and you don’t even need a RHN subscription to get the updates – they’re available too. Yet you’re arguing simply because you can’t download at no charge a branded copy of RHEL, this makes Canonical “FAR more open”? Sorry, it’s just nonsense: Red Hat do not develop proprietary software; every single thing they do is readily available and they even make all of it available free of charge.

  3. wariola

    Another Ubuntu Fanboy talking about Canonical puristness.

    Ok then lets talk about proprietary bits.

    Ubuntu Management Tools – Landscape (proprietary), no open source equivalent

    Red Hat Management Tool – RHN Satellite (proprietary), open source bits is called project spacewalk.

    Jboss Management – JON, open source bits for is called project JOPR

    Directory Server – Red Hat Directory Server, open source bits is called 389 DS (formerly Fedora DS) and Centos Directory Server

    Cluster File System – GFS and GFS2 is open source

    The only thing is that u cant use Red Hat branding

    And dont let me start on KVM, Libvirt, Virt-Manager, NetworkManager etc

    So tell me Canonical purist abt this?

  4. dogStar

    zoopster: “Canonical is quite transparent when it comes to what is open and what cannot be opened and EVERYWHERE it can, it opens – look at Launchpad.”

    Yeah, but look at Landscape.

    I’m not a fan of RH either.

  5. Alex

    wariola – actually, RHN Satellite is free software. The system is only open to Red Hat subscribers, but there’s not a huge amount of difference between RHN and Spacewalk now. Nothing Red Hat sells is proprietary in the sense of “closed source” (unless I’m missing something).

  6. This is a silly debate that is impossible to have with the business people at the companies involved (because we don’t care – we’ve moved on from this “I’m really pure and you’re not” debate long ago), but always involves those that don’t have direct stakes in the business. Give Michael a quota and he’ll start talking a customer’s language. He’d sound very different (which is the point I made in a zillion other posts, which Michael chose to ignore – customers care about good technology at a good price – they don’t spend more than a millisecond thinking about lock-in or they wouldn’t be buying Microsoft software en masse).

    The reality is that RHT does everything possible to lock in customers, within the constraints of Linux’s licensing. Why do you think they push the 3000+ apps certified so hard? To make customers feel happy? No, it’s to push the message that the water outside of RHEL is unsafe.

    Just look at its messaging on Oracle Unbreakable Linux.

    But why should we care? Customers certainly don’t. Not even the slightest shred of an iota. It’s only ideological purists that care about this debate. Red Hat gladly kept its Network bits proprietary…until it didn’t need to anymore. Again, who cares?

    I’m not sure why you choose to focus on the least important issues I raised in my post (purity!), to instead focus on the least important aspects of Michael’s (purity!). Michael is *very* worried about purity, but it’s unclear why he cares. Perhaps it keeps him warm at night. His customers, however, simply don’t care about his concerns.

    So why does he?

    It’s what I heard in Barcelona today from government IT officials. They spent an hour talking about how they want freedom in IT. I asked, “So why don’t you buy more of it?” Their response? “We buy the best software at the best price, regardless of the license.”

    That, my friends, is called “pragmatism.” It’s what the real world trades in.

    Red Hat sells because Red Hat often fits that calculus. Red Hat doesn’t sell because CIOs are sitting around in their salons talking about freedom with Diderot and Rousseau.

    Let’s get real. Ask your employer to give you a quota. The bigger, the better. Then you’ll learn to talk the customer’s language, and we can stop this silly ideological nitpicking over purity.

  7. Alex

    “Ask my employer”? I am my employer thanks, Matt, and to be honest it’s fantastically arrogant and insulting of you to imply that I don’t talk my customer’s language.

    But let’s move away from the snide personal remarks. You seem to be complaining that “purity” is the least important issue, but then front and centre in your article is this claim that “it’s not that businesses have bought into the ideological allure of freedom”. In other words, “assume my opinion is right and then we’ll talk” – sorry, no. I don’t agree that the “allure of freedom” is ideological; I think it makes cold hard business sense. That’s not to say it’s the only way of running an open source business, but let’s not pretend that freedom doesn’t have business value. It does.

    You do an awful lot of talking for Michael. His words are a bit different. When it comes to knowing who understands Red Hat’s customers, I think I’m going to side with Red Hat – they seem to be doing pretty well, so the empirical evidence is weighing on one side more than the other right now.

  8. Jose_X

    Matt, I think a part of the story you should not discard is that purity is important to many developers/contributors. If you want a large group contributing to your project instead of competing with it (fork or alternate base) or ignoring it then you want to speak the language of developers/contributors. In that language, available source is quite important.

    Of course, this importance goes down for a full-time developer that is getting paid well by a company that leverages closed bits, but this will not apply to all potential developers/contributors (just look at the competition and also note how many individual contributors fund themselves indirectly).

    However, since only one (or very few) companies will end up with relicensing rights to all/most of the bits of a particular source base, what you will find is that many of these companies that want these total rights will put their own bits out and compete against each other each with their own different bits. Except that it’s very possible that the largest subgroup among the competitors will be the subgroup that yields such rights. You’ll have company A bits, company B bits, etc. You’ll also have groups where everyone collaborates in exchange for each person giving up exclusive rights over all the bits. If the number of this last type of truly open project is one or a very small number, then it may very likely be the case that one of these will end up with the most contributors. Put differently, if I have to contribute to a company that insists on a special position or else contribute to a group where everyone has the same full access to all the code, all else being mostly equal, I’ll choose the latter. Most people will fall into this group, I think, because most people interested in a type of software won’t all work for the same company.

    Alfresco can tweak around with whatever business model they think currently is best, but they should remain wary that they might be blindsided by a new project whose maintainers attempt to be as open as possible. Openness of code will unite the largest number of developers over the long haul (all else being mostly equal). In the short-term, one project might really stand out and attract a large number of people despite keeping some secrets. Can Alfresco maintain a lead over a long period of time if they don’t open up fully?

    As for customers that don’t care about lock-in, many of these tend to adjust their opinion the longer they are exposed to FOSS. If the in-house devs find ways to add value that they could not add with closed source, then management will eventually take notice and adjust their opinion. [And devs love source much more than black boxes.] Remember that Microsoft is working from an incumbent position (huge market share) that developed during a time where FOSS was very low key. “FOSS” is still a little understood concept to many.

  9. Matt,

    Whatever your position, your discourse here is snitty and full of ad-hominem attacks. The same is true of your original blog post that Alex’s blog here references. Name-calling, as you have engaged in here, does not help, and it makes me far less inclined to listen to you in the future.

    Alex, I believe you provided a cogent, well-thought-out, logical, and correct analysis of the situation here. Thank you for that.


  10. Bernard Swiss

    Quote: “It’s what I heard in Barcelona today from government IT officials. They spent an hour talking about how they want freedom in IT. I asked, “So why don’t you buy more of it?” Their response? “We buy the best software at the best price, regardless of the license.”

    That, my friends, is called “pragmatism.” It’s what the real world trades in. /Quote

    No. That my friend, is called “expediency”.

    The use of this label frames mere short-term convenience as long-term, clear-sighted practicality, and dresses itself up in the more respected term, “pragmatism”, to avoid facing up to to inevitable but foreseeable grave consequences, and to discount the inevitable grave costs as being relatively minor and perhaps even unavoidable.

    The willingness to engage in this particular (self-deceptive?) name-game — ie. framing “expediency” as “pragmatism” — is a major and characteristic malaise of our time, and it is precisely this sort of behavior that led to the recent economic crisis, the decline of the US automobile industry, the Viox debacle, the ongoing failure to address climate-change issues, etc, etc. When are we going to learn, and start calling a spade a spade, again?

Leave a Reply