Italian Open Source developers: Alessandro Rubini (part II)

Alessandro Rubini, one of the most famous Italian hackers and representative person in the free software arena, yesterday answered few questions which get some local attention. Today I’m reporting the last of the two part interview.

How are you targeting businesses? Since you work alone, how do you cope with extra work?

Well, I actually don’t hunt for work. It’s usually the company that finds me. When a company looks for gnu/linux embedded expertise, they either go straigth to Montavista or they look around. If they look, they usually end up calling someone nearby. My recent clients are usually from my area, and that’s good: it means farther companies find other consultants (I know for sure they do). I’m not alone in my field, and demand is growing.

In this expanding working environment, I tend to cooperate with other consultants. We are actually mates rather than competitors, and it’s common for us to deal with overloads by getting help from a colleague, with due signing of NDA’s when the client requires it (not often, in fact). It’s not a nation-wide network, though. Strict cooperation is within a handful of people, but we heard about a number of others. So we can at least suggest other names to call when we can’t deal with the task at hand.

Alessandro’s network is the lightest form of association I can think of, a simple web page collecting people under the GNU Devide Driver umbrella. I’m happy to hear they take advantage of cooperation and I guess they don’t need to spend time and effort to write partnership agreements. It is a ring of trust, they all behave correctly and they naturally tend to do so.

On the other hand their approach doesn’t scale, they don’t share commercial costs and they are bound to spend time to commercialize themselves. They choose freedom, even in this respect.

Do you think Linux-embedded markets itself effectively to businesses?

It’s the other way round: many companies in the electronic and telecommunication area are seriously switching to embedded GNU/Linux.

Actually, it’s usually a little GNU and a lot of Linux (the kernel), but the reason they do that is usually the GNU GPL (freedom from lock-ups, more than cost issues).

Most of those companies just want to master the subject matter, or ask external help for the first project but work to build internal expertise meanwhile. I feel we consultants cover a very small fraction of overall investments in embedded GNU/Linux. But sometimes a company finds itself on strict deadlines and the internal resource reveal scarse; so they unspectedly have a tough problem for someone skilled to solve.

My suggestion, if any, to people willing to work in this field is to get expert in some field and publish the code they can contribute. This is the best way to get credits. Offering to help an established company (or consultant) nearby is a good bet nonetheless.
It’s quite difficult for us to find the right people, in a world where everyone claims to be a computer expert — so it’s easier for the right people to find us.

Alessandro feels that consultants cover only a tiny fraction of overall investments in embedded GNU/Linux, as he says his customers ask help just for the first project, but then they tend to internalize knowledge.

I suspect that many vertical markets are doing that nowadays, it’s definitely a resource marketing inefficiency, likely due to a sort of market failure occuring with imperfect knowledge.

How did your job changed during the last five years, and how will it change in the next five?

Things are getting more complex. As CPU power increases, companies want to put more stuff in it: graphic interfaces, video streaming and the like, even in small ARM or PPC devices. There’s less “substance” (industrial automation) and more “appearance” (cool gadgets), at least as an overall ratio. I don’t expect this trend to stop, and I’ll personally try to stick to “substance” problems as much as possible.

On a less technical level, it didn’t change a lot over time. There’s more work globally, as companies get more accustomed to free software ideas, but the basic points dind’t change much: they want to build their own expertise while delegating the first prototypes, and they come back when an unexpected problem occurs. Again, nothing disruptive is happening, it’s just the usual knowledge-related business practice. Which is a good achievement in itself, in my opinion.

Thank you very much Alessandro, you raised many important issues, I really wish you happy hacking, please keep writing books and educating new GNU developers!