What business are we in?

Print Friendly

Earlier today I was listening in on a meeting where one of our executives was making the point that people should stop thinking of themselves as being in the “open source software business” but rather just the “software business.” I want to riff on this a bit, but I want to make clear that I claim no originality in the thought above nor in how it develops below.

Which of the following statements are true?

  • There are companies that are completely in the open source software (OSS) business.
  • There are companies that are completely in the proprietary software (PS) business.
  • There are companies that are in both the OSS and PS businesses.

The third one is certainly true because, for example, IBM matches the description. We sell PS that (per the relevant licenses) contains OSS, we contribute to and are deeply involved in many OSS projects, we provide service and support for some OSS, and our services business involves both.

We are not alone, of course. Whatever else Oracle does with OSS, they have proprietary products that run on the OSS Linux, including that distributed by Red Hat. This is also true of many companies. Microsoft develops and distributes the proprietary Windows operating system but many, many OSS applications run on Windows.

So while I think we could find companies in the first two categories, increasingly we have vendors who have business models that involve both OSS and PS, intentionally or not.

Being in a situation where you have a proprietary product that allows OSS to run above it might generate revenue for that PS product, but might limit your ability to sell your own additional PS products that compete with the OSS. That’s life in the software business today. Various people and governments are examining these boundary situations between OSS and PS for both technical feasibility and legal compliance.

Therefore we have PS above OSS, OSS above PS, OSS in PS, OSS and PS more or less talking to each other, but we are not to have PS in OSS. That’s a legal point, but one that everyone should remember.

In the past we have had vendors who have proudly proclaimed themselves as being “the XML company,” or “the web services company,” or “the SOA company,” though I can’t recall if we ever had “the ASCII company.” That makes the point that such associations have a statute of limitations and it’s best to think of the next marketing slogan if you decide to do something similar.

I think in the same way we will less often hear about so-and-so being an “open source software company” and get back to the simpler “software company.” However, to the degree that anyone does advertise themselves as being in the open category, public and media scrutiny should examine exactly how much it is true.

It can go the other way as well. I recall a conversation I had with a company two years ago where the company was taking a very traditional intellectual property approach to standards and interoperability, and then told me that 90% of their products contained OSS. I thought, “Here’s a company in transition, even if they don’t know it.”

We’re all in transition, but some of us may not know it or perhaps admit it just yet.


  1. There’s an apocryphal story about a 19th-century “titan of industry” who was talking with his chief henchman about expanding into an entirely new business. “We are in the X business, not the Y business”, said one henchman. The robber baron replied as follows. “We are not in the X business at all. We are in the business of making money, and anyone who does not understand that is fired.”

  2. Maybe you can answer me a question. (And no, I don’t need ‘executive help’. I’m just curious).

    Every so often, I see an article like this one http://www.desktoplinux.com/news/NS6395384305.html , bemoaning that sometimes you can’t get various kinds of ‘Personal Computer’ hardware to work with Linux. I feel it too, because I have bought things like Microsoft Windows USB WiFi adapters, and they haven’t worked with Linux; presumably because there is some Proprietary Software required to control the thing.

    But then, I realise that we’re the same. I work at the lab which develops this http://www-5.ibm.com/storage/europe/uk/disk/ess/ess800/index.html ; slightly more expensive than a USB WiFi. In the middle of it is a ‘pSeries’ processor, controlling it, interpreting commands from the system it is providing storage services to.

    The ‘pSeries’ will need some software; and we’ve got two choices how we could do it. We can either make it run AIX and develop the device-specific software, and ship the package ‘Object Code Only All Rights Reserved’; or we can make it run Linux, develop the device-specific software; contribute the IBM-written software to ‘open source’, and tell the customer what level of Linux we tested it with.

    What we actually do, is to go the ‘AIX-Proprietary-Object-Code-Only’ route.

    Now, I’m reasonably sure that if someone buying a ‘shark’ wants to look at, or change, the adapter microcode, all they have to do is ask their account rep, and we’ll find a way. A Residency to write a RedBook, or an invite to the lab, or something. Especially if they are a big customer. Hire them and make them implement it, if they have a really good idea, maybe.

    So why ? If we’re going to be ‘commercially helpful’ to someone who thinks that having access to the source will help, why didn’t we open-source it in the first place ?

Comments are closed