Today I spoke with Tom Kemp, CEO of Centrify, creator of that remarkable patent-to-protocol map that I wrote about last week. My first question: why create such a thing? His answer: "Just the facts," and we went from that into a discussion of how open source and open standards suited his company and the market as a whole.
Tom's reason for creating the map was simple: He wanted to provide anyone who might be interested with hard details about which protocols might correspond to what patents. He also took the vital step of documenting each of his correlations, so that his work could be checked by others if needed. But making any kind of political statement about the results, so to speak, wasn't part of the plan and was the sort of work he would cheerfully leave to others.
The whole effort wasn't as difficult as it seemed, either, but a lot of the effort involved required that one know ahead of time where to look. "You had to follow what was happening and research the various documents, and once you were able to figure out where everything existed, then correlating everything took maybe 10 to 20 hours of work." If you didn't know the protocol names (and how they could be named different things in different places), the search would be that much more difficult. Tom said his experience with monitoring what was going on in the industry vis-à-vis the WSPP protocol issue, among other things, gave him a leg up in that department.
Tom was honestly surprised at how many people, me included, picked up on it and wrote about it. He was not trying to make inferences about the code itself, he stated; the code itself isn't something he has access to, and he can't comment on the code per se. "People seem to think protocols also equals code," he said, "but I'm not trying to map any of these to pieces of code; I'm just looking at what's been made public. With all that's out there about Microsoft and interoperability, there's been very little analysis of his kind, so that's what I wanted to provide."
A side note about Centrify: It creates solutions that allow Active Directory to be used to manage heterogeneous environments that normally don't use AD. My original guess was that Centrify itself had something of a vested interest in knowing more about the protocol/patent mappings, but Tom didn't think that was the case, at least as far as they did business. The Centrify products aren't themselves open source, but because they work with so many open source components, they provide a great deal of support for the open source ecosphere in their own ways. They sell support for open source technologies related to their product (OpenSSH, for instance), as well as create freely available forks of those technologies that can be downloaded and used in their environments. They also sponsor a number of open source projects with both monetary and programming contributions, such as Kerberos.
If anything, Centrify's business model is something like the MySQL model inside-out: The core product is closed source, but many of the accoutrements around it are open. In this case, the core of the product didn't need to be opened, but the APIs to it were kept as open and documented as possible. Tom drew an analogy with Salesforce.com, where the core is not open but the interfaces to it are. For him, a company going from closed to open source would need to ask: What needs to be open? What does the customer really want? If the core doesn't need to be open to get the most out of it, then it shouldn't have to be. Ditto the Mac: The core may be closed, but there's a fair share of open source surrounding it that people can connect with.
Despite being a strongly commercial outfit, Tom didn't rule out the possibility that Centrify could change business models in the future. In his purview, the whole question of open source business models -- and business models in general in this field -- are driven by what the customer needs. If a customer doesn't truly need something that's open, then it doesn't make much sense to go through a lot of effort to open it up for them. "There is no one right business model," he said, and went on to mention how there are also a multiplicity of delivery models, too: SaaS, appliances, standalone apps, etc. Everyone's trying to address their customers differently, and both MySQL and Ingres have found solid customer bases through each of their business models.
On a different note, we agreed that there are signs that Microsoft is de-emphasizing -- at least publicly -- its patent contentions with Linux in favor of a more friendly, we-can-work-together-after-all stance with open source generically. Whether or not it'll translate into any long-term changes for the company is another story, but with Ray Ozzie more visibly at the helm and the "Live Mesh" concept getting under way, it sounds like Microsoft is taking calculated steps from being a software company to a services company -- although probably not an open source company. Not yet, anyway.