Ten years ago, you might have thought hiring an IBM
employee to head up Linux development at the Open Source Development Labs
(OSDL) a strange move.
After all, IBM was once the poster child for a business culture that
emphasized conservatism, dark suits and white shirts; the roots of the Linux
movement, on the other hand, were found in the independent computer
enthusiast who cared more about writing great code than following a dress
As the saying goes, that was then and this is now. Linux is now firmly
entrenched in corporate networks around the world and lauded for its
stability and performance, while Big Blue has become one of the biggest open
source advocates in the software industry.
IBM — along with HP
, Computer Associates
and NEC — founded the OSDL in
2000. It has since become the home of Linus Torvalds, the developer who
first created the Linux kernel
exclusively on the Linux kernel while the OSDL focuses on enterprise
So it shouldn’t have come as a surprise when OSDL officials announced the
appointment of Tom Hanrahan, IBM’s Linux Technology Center (LTC) senior
program manager in charge of software development and testing, as director
of Linux engineering in late September.
Hanrahan will do many of the things he did at IBM, overseeing OSDL’s efforts
in helping the Linux community develop code that bolsters Linux in the
enterprise world, whether that comes in the form of technical support and
guidance for developers or equipment to test code on different
He talked with internetnews.com recently to discuss the challenges
ahead for Linux and some of his goals as director.
Q: When I think of the coordination needed to support the worldwide Linux
development process, I think of that commercial with the cowboys herding
cats on the open range. Any similarities?
Herding cats and organized chaos are kind of two sides of the same coin. I
would say this: development in the open source environment is significantly
different than development in a proprietary environment. What we’re finding
is that the strengths of working in the open source environment are pretty
exciting right now. And really, the key is you have so many people who are
able to contribute and who are able to review code that you can accelerate
the code development process while maintaining the quality of the code. To
me, that’s the real benefit that we get by having this new model.
As far as
how you manage it, you do have to think more in terms of how you influence
the direction things are going rather than dictating how things are going.
I guess that’s the sense you have when you talk about herding cats or having
Q: How much different was the corporate culture at IBM?
I worked at the Linux Technology Center [LTC] at IBM, so I was involved in the
exact same activities that I’m doing now. Part of the training that IBM
gave us before they brought any of us into the LTC was that you do have to
understand how this different model of software development works. IBM did
a pretty good job of preparing us and training us for what to expect. So
when we started entering the community, we entered in a way the community
would expect developers to enter, and we didn’t come in with a proprietary
mindset. I think that’s what made IBM’s participation in the community
Q: What are your first impressions of the OSDL?
My first impressions are that it’s confirmed the good work I thought OSDL
was doing all along. That they really do provide value to the community in
terms of human resources, such as developers writing code and testers
providing test information to the development community, as well as
equipment resources so that if people have a need to test on a particular
architecture, a good place to look is the OSDL.
Q: When you’re sitting at your desk, do you ever get the impression, “oh
man, Linus is looking at me,” the sense he may be watching and seeing what
you’re up to?
[Laughs] Not so much. I’ve met him a couple of
times and he’s a really good engineer, but I’m used to working with really
good engineers. So I don’t feel intimidated by having him being a part of
Q: What do you see as some of the biggest challenges for Linux in today’s
I think one of the things OSDL can do is to help better integrate the ISV
development process with the Linux community development process. I’ll give
you a specific example.
ISVs tend to work in a cycle that’s six months to a
year. They expect to be able to have a stable platform underneath them in
order to recoup the amount of money they’ve invested during that development
effort. The Linux community is at a much more aggressive development cycle;
there are some times when the code is updated daily; weekly updates are not
unusual; [and] monthly updates are common.
We have to find some way to help ISVs work in that kind of environment, and
one of the initiatives we’re starting right now is something called the
“binary regression test.” Our thought behind that is we would work with
the ISV to see what kind of binary regression testing they do as part of
their six-month to one-year cycle.
We would create a test suite that
mimics that regression test suite, and we would begin running that on a much
more aggressive schedule so that we could quickly identify
any break in binary compatibility so that we can work with the
community to fix that break or we can better prepare the ISVs for what
they’re going to have to do when they move on to the next version of Linux.
On the desktop side within the enterprise, that’s one we’re still learning
what the challenges are. The biggest challenges I see are how
we can better co-exist with proprietary operating systems, particularly
Windows. The more that Linux can work in a mixed environment, and the
better that Linux can work in a mixed environment, the greater its adoption
rate is going to be.
Q: There are some who are worried Linux might be chasing the tail of
Windows functionality. At what point do you decide which function is good
for Linux or is just a Windows copycat, or any other operating system?
We don’t make those decisions at all. It’s really the Linux community that
decides the direction it wants Linux to go in. In general, there’s such a
wide variety of people who use Linux that the right thing happens for each
usage at every step along the way. For example, the graphical user
interfaces right now for Linux are really pretty good; someone coming from a
Windows environment to work on GNOME or KDE is going to be pretty
comfortable with how things work. And these were just decisions that are
made by the community as they start wanting to use Linux themselves as
Q: Have you set any goals for where you want to see Linux two years, five
years down the road?
I’d definitely like to see us continue to grow on the server side. The Web
server market, for example, is pretty well penetrated by Linux right now.
The telco market is becoming much more comfortable with Linux, and I think
there’s pretty good penetration of it there. I’d like to see Linux make
advancements in the data center, and there are a number of markets, like
financial services, where we’d like to see faster adoption rates.
On the desktop side, I’d like to see us have better support for the
independent software vendors; I think that’s the big challenge we’re facing
on the desktop side. Getting a big pool of applications available is a
challenge, because it’s still pretty hard for ISVs to kind of pick the
combination of Linux and GUI that they want to target.