Getting Open Source Right
May 23, 2008, 22:30 (6 Talkback[s])
(Other stories by Brian Proffitt)
Full Text Search: The Key to Better Natural Language Queries for NoSQL in Node.js
In his rationalization about why government-mandated open source
software deployment is not a good idea, Microsoft's Director of
Standards made some assumptions about what the real value of open
source software (OSS) is for governments.
Like most assumptions about open source from Microsoft, a big
piece of the overall picture is conveniently left out.
The context of Matusow's
blog entry is a trip he recently took to South Africa where he
had talks about interoperability and open source issues. As noted
in a Tectonic article, Matusow's visit fell around the same time
the South African government was approving the OpenDocument Format
(ODF) for government use. (Funny...)
Matusow argues that the mandate of any one IT methodology is
really not a good idea, because it tends to lock-in governments to
only one way of doing things. Mandate only open source, he argues,
and then you leave out all of the other nifty tools you can use for
Which, on the surface, does seem to contradict the whole notion
of freedom that OSS advocates cherish. But if you dig just a little
deeper, you can see there's some shaky assumptions propping this
The first assumption is that governments who choose OSS are
trying to generate a strong local economy/development ecosystem.
Matusow then generally claims that this kind of ecosystem is
improbable to create because any real innovation in the open source
world is just going to be snatched up by the few big commercial
vendors anyway, such as Red Hat. This assumption basically paints
the entirety of the open source collection of companies as a single
entity that has just as much of a chance to suck resources and
create vendor lock-in as any other single vendor.
This kind of altered reality is really the only way an OSS
lock-in argument will work. Do the commercial OSS vendors
incorporate home-grown technologies into their final products? Of
course they do, that's the nature of open source. But Matusow
blithely ignores the fact that any software used and changed by a
big vendor will have its changes pushed back upstream to the
original, local coder. He also ignores the reality that the
availability of large OSS projects will allow the local companies
to build their tools faster and better than they would completely
on their own.
And then there's the notion of OSS vendors as one, big entity.
There's an entire spectrum of OSS companies, all happily competing
against each other. If Red Hat were to try to make a deal with a
government, don't you think Novell and Mandriva would be there
trying to make the same deal? Competition, to me, means that
customer has more choice, not less.
Matusow then goes on to make another, more specific, assumption
about South Africa. Their local developers don't seem to have the
skill set to do any real contributions to OSS anyway, so the
benefit of local participation is lost on them anyway.
Ignoring the fact that this is the single most arrogant thing I
think I have heard anyone say about a nation's potential (I'll let
colleagues take that part of the argument on), who is Matusow
to say anything about who can code what for Linux and OSS? Is you
said "nobody," you're absolutely right.
Matusow cheerfully tells his blog audience outrageous statements
about South African development skills, all the while missing one
of the biggest parts of OSS development: it's an open meritocracy.
If you or I want to code for Linux, we can look at the code that's
there and learn from it. If we learn enough, then we can try our
hand at sending in some code to some small part of the kernel. If
the lead developer for that section likes it, maybe it will be
included in the next merge window. If Linus or one of the lead
maintainers likes what we wrote, maybe it will be officially a part
Note, that at no time in this scenario, did anyone ask anything
about where the developer is from or how he learned to code. No
one's checking passports at the door. Ever. Contributions are based
on how good the code is. I realize that's a hard thing for someone
at Microsoft to grasp, but anyone who claims to have watched open
source development should know that by now.
Is this always the case? No. Sometimes errors creep in and are
not caught effectively. Sometimes personal disputes between
developers will keep a nice bit of code out of a project because
someone's having a fit of pique. But in general, the practice of
Linux and most OSS development projects is to take the best code,
no matter who built it or where it comes from.
This isn't just because Matusow, a Microsoft employee said this.
I should point out that it would be the same case if someone like
Linus Torvalds or Richard Stallman or Ian Murdock went insane and
made the same statement. None of these guys, or any other player in
OSS, big or small, has the right to issue sweeping generalizations
about who's worthy to participate in the development of a free or
open source project.
That's what open means.