Interview with Linus Torvalds

I can not begin with this question: is the Linux kernel source code really violating any patents of Microsoft?

As far as I know, no, and this step by Microsoft is just an attempt to slander its rival, in competition with which at the technical level they have some problems.

According to Mark Shuttleworth (Mark Shuttleworth), the most important feature of Linux distributions is that they are free. He said that this system would die if the whole world stopped buying the Vindouz packaged in a gift box and started buying the Linux packaged in the exact same box. How can you comment on this statement?
')

Yes, I totally agree with him. And I do not see that Linux is experiencing any problems in this regard. It seems to me that if companies, including Microsoft, work together, there is nothing wrong with that. But Linux itself has nothing to do with the recovery of patent royalties. In fact, the open source license agreement (GPL), in its second edition, already assumes that the software can be freely distributed without any patent restrictions whatsoever.

What do you think of the third edition of this license?

For me, this is just one of over fifty other open source licenses, such as BSD, MPL, and others. It’s not as bad as the previous draft versions were, but it seems to me that GPLv2 is simply better.

New editions do not mean at all that the license has become better, especially when they are developing in the direction of complication and impose much greater restrictions on the consumer.

You are now a member of the Linux Foundation consortium. James Zemlin, director of the organization, said in the New York Times: "There are things Microsoft does well in promoting Windows OS, providing legal protection and standardizing system components." He also said that "what Microsoft does well, we must do well — promote, protect, and standardize Linux." In your opinion, what else can we learn from this corporation?

Historically, the main lesson of Microsoft, which they themselves seem to have forgotten, is: "Give your customers what they want."

It seems to me that the reason for the great success of Microsoft is that they filled a niche with some basic technology (in this case, I’m talking about their BASIC language - it was thanks to him that they received a big impetus in development), and they were selling technology made on “good enough "Level at a low price. At the time, they did not play games with their consumers.

Of course, since then a lot seems to have changed. For the past few years, they have been playing with their users: the seven versions of Vista and the hard-coded DRM are clearly not what everyone wants.

Microsoft has always had a good sales policy. Their strong market position also contributed to the standardization of Windows as a platform. By and large, this is good for the user. Now they are moving away from the old achievements, for example, they are promoting the new Vista to the market and providing Direct X technology only for it. I think their historical success deserves attention.

What do you think of the Novel and Microsoft agreement? What will this union bring in the future? And what can you say about the events in Redkhet?

I don't care about all this. You ask all these questions about companies and marketing, but I have nothing to do with it. I am absolutely not interested in this. All I'm interested in is technology and work with people.

In the era of Web 2.0, we have a great understanding of the open source software development model, the Linux model. I think the Adobe companies, in part, Microsoft and San, too. What is open source software now?

I think the question of accepting an open source model is that no one really can design a complex system. This is simply contrary to the natural course of things: people are not smart, not a single person. The open source model, however, allows not to design things, but allows them to evolve, passing through the barriers of the consumer market. Thus, the end result is constantly improving.

And doing it openly, allowing everyone to interact with each other, share ideas, without restricting progress with nondisclosure agreements - this is the best way.

I can compare it with science and witchcraft or alchemy. Science may take several hundred years to understand how the world works, but ultimately it will achieve results because people can share knowledge, they can develop theories of other generations, and their knowledge evolves over time. On the contrary, really intelligent people can practice witchcraft or alchemy, but their knowledge is not accumulated anywhere. They, of course, can pass on knowledge to students, but data hiding does not usually lead to an improvement in the results that a particular person has achieved.

The same can be said about the opposition of open and closed software. People may invent something really valuable, but at one point it will become prohibitively difficult for one person or the whole company, and the personal goals of this person / company will always limit the development of the product.

On the other side is the world of open source software, which copes with complex, complex solutions. No one individually can take the whole picture as a whole, but development is possible without such a solid perception, only local improvements and an accessible open market are sufficient.

Thus, it seems to me that many companies are gradually beginning to adopt the open source model, simply because they see that it works, and they realize that it will be hard for them if they want to play it all on their own.

Linux is a multipurpose system. It employs personal computers, huge servers, cell phones and many other devices. From your privileged position, tell me which sector of Linux devices will be able to show the greatest potential?

The power of Linux, in my opinion, is that it is not intended to occupy any particular niche. Everyone can meet their needs, and many people, many companies have completely different goals and opinions about what is important to them. So I'm also not interested in any particular Linux application.

Personally, I am inclined to believe that the system can show the greatest potential on home desktop systems, not because this is our primary goal, but simply because desktop systems assume more diverse and complex behavior than any others.

In other words, for desktop computers, the system needs to be more advanced. On servers, there is usually no place for such a variety of software and hardware. On many embedded systems, the system should really do only one or two tasks well.

If we talk about desktop computers, then different people perform different tasks, and the system needs to be able to perform any task correctly on any set of equipment.

People are still waiting for the magnificent entry of Linux on desktops. Distributions with friendly user interfaces, such as Ubuntu, Dell’s decision to sell computers with Linux preinstalled is undoubtedly two big steps forward. But it seems that something is still missing. What do you think?

For me, it's just a matter of time. We already have all the necessary pieces of mozayka, we can improve each of them. Simply, the inertia of most people and companies who do not want to give up their familiar surroundings for no particular reason manifests itself.

Therefore, I’m not particularly worried about something specific, I care about Linux being gradually improved, and time will do the rest.

Eben Moglen asked Google for more help from the open source community. How do you describe your relationship with this company?

In fact, a lot of kernel developers work on Google, so I wouldn’t worry about that. My indispensable assistant Andrew Morton (Andrew Morton) works for Google, and its immediate goal is to improve the kernel. And this is all that is ultimately important: even large companies are just a group of individuals, and it’s not important whether Google helps specifically, but that people like Andrew work with us. By hiring such people, Google ultimately helps us.

And, of course, this does not apply only to Google. This is true for any company that is somehow involved in our work, and if not involved, it allows its employees to devote time to developing open source software.

A little technical question: any news about the future of the kernel? Perhaps plans for branch 2.8?

We are not planning to switch to a new versioning scheme for the kernel: we have achieved great success by applying the current development model, releasing a new kernel version 2.6.x approximately every ten weeks (2-3 months), and since then we have managed to implement quite serious changes no less having good reason to change the current version of the kernel number greatly.

And, in my opinion, it should be so. Smooth and lasting improvement. Over the years, we have carried out many serious reorganizations of the kernel code, but since the kernel has matured (and we finally understood how best to maintain it), we have fewer and fewer reasons to make drastic changes, and more reasons to see the process of kernel development. like systematic improvement, which most users don’t need to know about.

Speaking from the user's point of view, he does not want to switch to the XY core, since it knows this and such a thing, he wants to be sure that he can just upgrade his core, and things that worked well will work even better.

This may not be particularly entertaining, but it is reliable. In the end, kernel developers find things interesting that the average user doesn’t think about (and shouldn’t think, because the operating system should only be a layer of abstractions between system resources and applications that you run on top of these resources).

For fun, which distro is your favorite, and which one do you find safest?

I really do not really care. For all the time, I changed a lot of distributions, and for me, the most important things in a distribution are installation simplicity, ease of updating and availability of kernel changes — this is especially important for me.

One of the popular distributions that I have never used is Debian, precisely because it has traditionally been difficult to install. It sounds a bit strange, because Debian is considered a purely technical distribution, but this is literally something I don’t want from a distribution. I prefer those that are easy to install, because this is exactly what you have to do with the system in the first place.

I used Susa, Redhet, Ubuntu, Yellow Dog Linux (I had to use the PowerPC architecture, and this distribution was a good choice). Now Fedor's seventh is installed on most of my machines, but this is just a fact, which does not mean that Fedor is better than other distributions.

In a famous book, you defined the development of Linux as an activity "just for fun." Do you still enjoy the development?

Yes of course. This is the reason why I keep developing. At different times, the work I enjoyed was different. Previously, it was mostly coding, but now I do not write so much code, but rather do organizational work: redistribute parts of the code, communicate with people, direct people to the right direction, and sometimes correct my own mistakes.

______________________________
Original interview in English.