Interview Linus Torvalds

zhaozj2021-02-08  311

Interview Linus Torvalds

Interview

Alessandro Rubini Translator

Zhao Wei

(This article is reproduced from Software Engineering Expert Network

Www.21cmm.com)

Alessandro: Can you take more time to work for us?

LINUS: Of course, no problem, I will do my best.

Alessandro: Have you been in the United States? I prefer our Finland leaders, rather than a novel United States .... How do you see this problem? Consider returning to Finland?

Linus: I am very good at this now: I really like the days in the University of Helsinki, but I think some different things is also worth it, and everything is very good so far.

I agree that Finland is more "neutral" in many ways, which is very beneficial for Linux: There are many people who are sensitive to the United States or even don't like the United States, but I really don't know who is really don't like Finland. So it seems a little bad, but I feel that everyone trust me, not especially paying attention to me is a Finnish, so I don't think this is a very serious problem.

The weather here is much better than Finland (although this is the most rainy winter in the Gulf area), life here is more interesting. Although I really like my work in the university, I am doing something is more exciting, soon it is going to be completed, say so much.

Alessandro: We all know the US restrictions on encryption technology; have they affected Linux development in this field?

Linus: This seems not particularly serious. Before I moved to the United States, the US stupid restrictions on encryption technology were already a problem: mainly because if we want or when we want to join the support for encryption technology during the kernel, we must not Without separating this module, this CD business (most of the US) can sell their CD overseas. So I moved here for this matter, I didn't have a real impact. I couldn't be able to do my encryption now (in fact I originally not responsible for this work, I am not an expert in this area) .

But no matter what, I still hope that the US encryption method will be quickly changed. Some of the United States has made a lot of complicated things that have been simple. (Fair thinking that the United States is not the only country: France is more serious in this area, and they are trying to make the entire Europe like them. Fortunately, the french crazy ideas do not have much effect. The United States is different, and they are lively in the field of information technology).

AlessAndro: Have you ever thought about giving up your Linux coordinator? Is this job interest like this? If you plan to leave, can you tell us about your next plan?

Linus: I have never really want to leave, have someone asking me if I am not interested in this, then what else is there? Linux coordination is always very interesting, although I always think it takes me too much time.

Alessandro: Please forgive my curiosity, how much time you spend now every day? What is your main activity now?

Linus: Now I have not so much time to compile: I still have a code for several weeks a few days ago, but I'm so doing so, most of them is because the kernel needs to do some basic changes. . Last year, there were four to five times, most of which were because SMP, and the cache of this "Dentry" file system.

Most of my time spent reading email and reply - coordinating everyone's work, talking about my own point of view, organizing PATCH. Now this is most of my work: confession, programming only 10%, coordinated work accounts for 90%. Alessandro: How did you handle this thing? On the one hand, write a free kernel, while still can earn money?

LINUS: At the beginning, I am a student at the University of Helsinki. In Finland, this means that in a few years, you can get the government's funding to help you complete your studies, and you can get special student loans. I guess the situation in Italy may be similar to this, but maybe unlike Finland.

One or two years later, I started to help my teachers in college, and then the assistant researcher, the school was always encouraged to continue to write Linux.

Now, I am obviously working for a business company, but even here, I still have to do a lot of Linux work, even in work, although Transmeta's business and Linux have no relationship, a large number of internal uses Linux, so my Linux work is still good for the company.

So, whether it is working in college or learning is good, or working in business institutions, I can always "go to work" and Linux. Never have a serious conflict, although my work hours are not strict restrictions on 9:00 pm ...

Alessandro: Why don't you try to turn to commercialization? Just like Cygnus? (I think I know why :-)

Linus: I have never turned any part of Linux to commercialization: maintaining a commercial company or a similar institution, will spend a lot of time, I am not interested in this.

And this will bring a burden: I can't do what I want to do, I have reasonable things, because I have to consider that I have to make money to feed me and my family. In contrast, in the university or in Transmeta, I don't have to rely on Linux to earn money to support myself. In this way, for Linux, I am completely free, I don't have to worry if it will send me the next month's salary ...

I am very happy that I don't have any pressure on Linux. I think most other developers are also related to this problem (they don't have to worry about my judgment is affected by any fiscal issue).

Alessandro: Do ​​you think you have changed the world? Or just a first step? (Oh, I know you)

Linus: This is from me, I feel very proud. I don't think I "change the world", but I am very proud, I have changed many people's lives - this feeling is very good, when you realize that there are many people concerned about your work.

Of course, I will not say, "makes my life meaning", but Linux is definitely part of my life, if you understand what I mean.

Alessandro: How do you see the work of Richard Stallman, and his philosophy?

Linus: I personally don't like to stir together political and technical issues, and I don't always agree with the RMS in many problems. For RMS, when it comes to software, there are many problems that are almost religious beliefs. I am more realistic in most issues. So, we know that our views are different on some problems, we don't try to work together because we know that this will not be smooth.

What is said above may make people think that I don't like RMS, it is not entirely the case. RMS is obviously most of the "free software" or "open source" trend, if the RMS is missing, the world will have a lot of boring. And in order to promote this trend, he "must" appear to be partially boused. So accurately, I really admire RMS, but I don't want to be like him, because our world outlook is different.

Alessandro: How is the problem of reality? What is the schedule of 2.2? 2.0 and the main difference between 2.2 on the horse?

Linus: I have seen it now, 2.2 will be issued earlier in the summer, but it is difficult to say: there are some things must be modified, and they are not allowed before correcting these issues. At the moment, this TCP card has a good job, and the other parts are very good, but this problem is very serious, I still don't want to think about 2.2 before it gets modified.

2.2 One change is that some of the new features in 2.0 will become more mature in 2.2, such as SMP, and Multiple Architectures. There are still many new things (new Dentry code, completely rewriting NFS, etc.), but SMP and Architecture will be one of the most basic things in 2.2.

Alessandro: Bruce Peens claims "Occupy World: 2003"; Is this actual? According to your point of view, will the concept of free software win the market? In this regard, how do you think about the change of the Netview Company?

Linus: "Occupy World", it seems that everyone is saying this, but I think, yes; if there is five years, it is not realistic. The open source of the net view is a good start, I think there will be other people soon.

AlessAndro: How can various free operating systems CAS? What is your point of view?

Linus: I think everyone is now realized that others have existed, but there is still no cooperation between this mutual cooperation now. As for Linux, it can work together with other other operating systems (especially Microsoft's very bad operating system;), I didn't see the signs of change, FreeBSD also has other systems. Perhaps some changes must be made on some details, but I can't see any fundamental problem.

Alessandro: You think Wine projects, there are some other tools that can make two technologies, but a free and free, the other is all property rights, the operating system cooperates, run the same application? (Terrible issues, my personal opinion is).

Linus: No, I think the Wine project is a very important step for the operating system on the PC, but this step has caused a wrong impression: it seems that a good operating system must be able to run those Windows applications, such as MS Office, etc. Otherwise, it is not a good system.

So, I think each system should be able to run some basic procedures, but there is no need to ask two systems to completely equivalent, they have their own different focuses. For example, Microsoft always emphasizes performance as long as they have to go, and their volume is very large, and Linux is (and will always be) emphasizing technical issues.

Alessandro: Currently we have free office applications. Do you think this is just a problem with sooner or later, or this software is only commercialized? Linus: I want commercial software to have a place, although I think we will have free office applications soon, but I don't think we can't do this.

I want a free free operating system and some basic applications. I think if these basic things are not strong enough, and you can't modify it freely to suit your needs, you are really encountered. It's troubled. But talk about other fields, this point is not so correct, and you can get the source code for free, nor is it as important.

Alessandro: Sometimes we will notice some "standard" is still all of the property rights (such as I20), is this just the dying struggle of these companies, or saying that the free software is actually in danger?

Linus: I am not very worried about I20 and other standals of some property rights. All the standards of property rights, this idea has failed - almost all successful standards are open. Sometimes, all property rights are very exciting because of these standards, they are not willing to give up, I guess, in addition to Intel and Microsoft, there is no other person, even the two companies will be in competition. It is weakened.

Alessandro: What is your location in this matter? That is, you can only get the Linux module in binary.

Linus: I have accepted this extent, but I never supported, and I don't like this.

The reason I accept it is that in many cases, you have one, more than the side of the device driver written for Linux, which can work in Sco Unix or other systems, at this time, the equipment manufacturer suddenly awake Come, realize that Linux is more promising. The result is that they transplanted their drivers to Linux.

But because this driver is not "based on" Linux (it exists independently of Linux), I don't think I have the right to ask them must be jurisdiction from the GPL, so the module interface of these non-open source allows these modules to be in Linux. Lower survival, work.

This is not to say that I will accept any module of blocking source: Some things are obviously based on Linux, and it is not straight to the Linux kernel. In these cases, this is clearly a job that is not completely original, and it is no longer applicable, which is clearly the scope of the GPL certificate.

Alessandro: What do you think about this problem with KDE-QT? Will GNOME will succeed?

Linus: I personally like QT, and KDE looks good. In this whole thing, I took the attitude of walking and see if GNOME can do very well ...

Alessandro: An interesting challenge in the network subsystem is "Band RESERVATION"; how is Linux?

Linus: I have to have to skip this problem. I personally didn't participate in this area, and I didn't be particularly interest, so I would not spend too much energy on this. This is the way Linux work: If you need this, then you do this work, if there are many people need this, then we integrate it into the system ..

Alessandro: There are many people asking why C write system kernels with C instead of C . What is the reason why you use C in the kernel? What is your favorite language, except for C?

Linus: C allows us to use some of the characteristics of the compiler I like, and in fact, we used C in a short period of time, just before publishing Linux-1.0. As a result, it seems that this is not very useful, I think we will not try again, there are several reasons. One reason is C more complex, the compiler always makes some things that make the programmer unexpected. Of course, you can avoid these features such as Virtual Class, but C allows many C to be allowed, which makes it difficult to debug, modifications.

Another reason and the above reason are also very relevant, that is, the speed and reliability of the compiler. Because C is more complicated, there may be bugs that may appear, and it will always be slower. It can be considered that this is just a compiler, but the complexity of C is objective for the development of kernels.

Alessandro: How do you see the Java phenomenon? Have you considered integrated Java virtual machines in the kernel? Such as Kaffe?

Linus: I always feel that everyone is too enthusiastic to Java, and now I am still like this. My sincere "hope" Java can succeed, but I am very actually, I will not join Java.

Linux has long supported the seamless operation of the Java big program. I think the Linux kernel is just as a parcel, not a direct Java virtual machine, which is an advantage.

This article is reprinted with the permission of Infomedia, Italy. An Italian translation of this article can be found at http://www.pluto.linux.it/journal/pj9807/linus.html. The interview was done by e-mail IN february, 1998.

Excerpt from Linux

转载请注明原文地址:https://www.9cbs.com/read-604.html

New Post(0)