25 years later: an interview with Linus Torvalds

The first issue of the Linux Journal published an interview taken by Robert Young, the first publisher of the magazine (and, among other things, the founder of Red Hat), by Linus Torvalds (author of the Linux kernel). We decided that it would be interesting to bring them together again after 25 years. The first interview can be found at the link .

Robert Young : The occasion to contact you was a great pleasure for me. How are you, like family? Your children have probably finished college. Nancy and I have three daughters, everyone is fine. The eldest, Zoe, who was 11 years old when Mark and I started the Red Hat project, will soon be born second - that is, I am already a grandfather.
')
Linus Torvalds : My children haven't finished college yet, although Patricia (the eldest) finishes in May. Celeste (the youngest) studies in the last grade of school, so in six months our nest will become empty.

All three have everything in order, and I hope and suspect that in a few years, when the whole story begins with the grandfather, everything will also be in order.

Bob : When I interviewed you for the first interview in 1994, did you think that you would still support this thing in 2019?

Linus : I think that by 1994 I was already surprised that my last project did not become another project in the series “do something interesting until it does everything I need and then find something else.” Of course, the development was at an early stage, but the project has already healed with its own life.

So I think I didn’t expect to do this for a few more decades, but the project has clearly already moved to the stage where it turned into a fairly large part of my life. I didn’t have long-term plans for Linux, and I’ve just dealt with them as they are developed, without worrying about what will happen in five or ten years.

Bob : There is such a famous old saying about the danger of achieving your dream - your favorite joke at the time when you said that the goal of Linux for the future will be “power over the world”. And what's next when you and your open source and free software community reach your goal?

Linus : I stopped joking about the power over the world for a very long time, because over time this idea became less and less comic. But it was always a joke, and everything that I and other developers did, we did not for this reason. Motivation has always been the improvement of technology and the solution of interesting problems.

And, in fact, nothing has changed. All the details have changed - the hardware is different, the problems have become different, my role has changed. But the principle of "do better and solve interesting problems" has not disappeared anywhere.

For example, in 1994 I was mainly engaged in the development. I was in charge of the project, but, although I spent a lot of time on combining patches, I mostly wrote the code myself. Today, I rarely write code, and if I write, then it is pseudo-code or examples of patches that I send to real developers. I would not call myself a manager, because I do not do annual reports or budgets (and thank God), but definitely a more project manager than a programmer, and this situation has been maintained for many years.

So the whole picture has not changed, but my role and all the details obviously looked completely different in 1994.

Bob : What will happen to you and this base of the code in another quarter of a century?

Linus : Well, then I will be 75, and I doubt that I will deal with such issues every day. But given that I've been doing this for 30 years now, maybe I will still follow the project.

The good news is that we already have a good developer base, and I'm not worried about questions like “where will Linus be”. Yes, people have long been arguing about how kernel developers are getting older, but this is not due to the fact that we will not recruit new people. It's just that we still have people in the team who have been with us for a very long time, and they still like to do this.

I used to think that one day there will be some radically new and interesting OS that will replace Linux (yes, in 1994, I could afford to think that maybe Hurd could do it!), But we are not just doing this for a long time and successfully, I also understood that the creation of a new operating system is a task much more complicated than it seemed to me. This really requires hard work of many people, and the strength of Linux - and open source in general, naturally - is that you can act based on the work of all other people.

So, unless some kind of tectonic shift occurs in the computer landscape, I think Linux will feel good for the next 25 years. Not because of some features of the code, but fundamentally, because of the development model and the task area.

I may not be working so actively, and a lot of code will be updated and replaced, but I think the project will remain.

Bob : Have you and your team been satisfied with updating kernel code all these years? Is there a need to rewrite some part of the ever-expanding code base, recruited over 25 years? Perhaps in some more modern dialect C?

Linus : During these years, we have rewritten our subsystems a lot of times - of course, not all at once, and no one wants to change many parts of the code (most often, this is a driver of outdated equipment that few people use, but we support it). One of the advantages of a single code base for the entire kernel is that when we need to make a big change, we can do it. Of course, there may be all sorts of drivers and other programs in the out-of-tree format (both in source and binary), but we always had a rule that if something is not included in the main tree, it does not matter for development . Therefore, if necessary, we carry out radical changes.

As for C, nothing better has appeared yet. We updated the source code of the kernel code for improved language features (the C itself has also changed over the years), and added different extensions over C for additional type checking and checks during program execution, etc., but in general the language is almost the same as except for small details.

And, apparently, hardly anything will change. Programming languages ​​that are not intended for low-level system programming are being actively developed. They are needed to facilitate the creation of custom applications with a trendy interface, etc. They specifically don’t want to do what the kernel needs, such as low-level memory management.

I can provide a special “platform” for creating drivers or something similar, and we have a simplified “language” for configuration inside, and we use some languages ​​for assembly, so it’s not that we use only C. But for the most part it is C, and the core is written on it.

Bob : What is your favorite development computer? Are there any Linux laptops that would be something like Stradivari's violins for musicians? Or tablets, or phones?

Linus : My main development machine is quite an average PC. This franken machine, assembled over many years from different parts. Nothing special, and the last time I upgraded it a couple of years ago, so nothing new. My main request to the computer is complete silence. In addition to a couple of fans, there are no moving parts (no more rotating disks), and most of the time the fans are turned off.

On the road (which, fortunately, is rare), I usually need a good screen and light weight. I strive for weight in 1 kg along with charging, and so far I can not achieve this ideal, but so far my best compromise is XPS13.

Bob : It seems that the main success of Linux in the field of desktop computers was not in the PC world, but in the Android world. What do you think about it?

Linus : Traditional PCs have already lost their dominant position. Even if a person has a PC (and even if he is still running Windows or OS X), many still mostly use a browser and a couple of random applications. Of course, there are “workstation users”, something like a desktop computer that I always imagined. And, despite the still continuing importance of this role, it already, apparently, does not steer the market, as PC used to do. Powerful desktop computers are now only needed for development or games, as well as audio and video editing. Typically, a computer is used to launch a browser, and more often it is just a tablet or phone.

Chrome is doing well in this area. But, yes, if you simply count by numbers, then a large number of people who deal with Linux every day are Android users.

A remark from Bob: If we consider “dominance” in the strict sense, then this is probably true. But, despite the recent decline in PC sales, the cumulative growth of the personal computer market from 1994 to 2014 is so powerful that even today’s slower market, 4-5 times more PCs are installed in the world than in 1994.

Bob : If you had to fix one thing in the networked world, what would it be?

Linus : Nothing technical. But I just hate modern “social networks” - Twitter, Facebook, Instagram. It's a disease. She encourages bad behavior.

I think that in particular this is also present in e-mail, and I have already said once: “Nobody catches your hints on the Internet.” When you don’t talk to a person personally, face to face, and skip all the usual social clues, it’s easy not only to miss the humor or sarcasm, but also to miss the opponent’s reaction, which causes things like flame, etc., what does not happen in person

However, email still works. You need to spend efforts to write a letter, and it has certain content, technical or some other. This whole model with “likes” and “share” is rubbish. No effort, no quality control. In fact, everything works as opposed to quality control - click-bates, things sharpened by an emotional response, etc.

Add anonymity here, and get something disgusting. If you don’t even sign with your name this garbage (or garbage that you share or like), this does not correct the situation.

I am one of those who believe that anonymity is overrated. Some people confuse privacy with anonymity, believe that they are interrelated, and that protecting privacy means protecting anonymity. I think this is wrong. Anonymity is important to informants, but if you cannot prove that you are who you say you are, then your crazy chatter on any social platform should not be visible, and you should not be able to like or share it.

Well, okay - chat on. I am not in any social networks (I once tried G +, because it didn’t have this ordinary brainless nonsense, but it came to nothing), but they still annoy me.

Bob : The current issue of the Linux Journal issue is Kids and Linux. Would you recommend something to young programmers and computer science students?

Linus : I should be the last to be asked. From an early age I was interested in mathematics and computers, and before the university I was self-taught. And everything that I did, I did because of intrinsic motivation. Therefore, I do not understand the problems of people who say “what should I do?” This is not my topic.

Bob : We first met at the Digital Equipment Company (DEC) fair. This was your first trip to the USA sponsored by John "Mad Dog" Hall and DEC.

Linus : I think it was my second trip to the USA. During the first, like, I went to Provo (Utah), to talk with Novell about Linux (about the internal project of Novell, which later turned into Caldera).

But, yes, the DECUS fair (in New Orleans, if my memory serves me) was among my first trips to the USA.

Bob : I asked you then how will you deal with all email when you return to Helsinki. Your answer surprised me and since then I quote you. You just said that you would send all the old letters to / dev / null. I was shocked and asked you, “what if there are some important letters in the mail?” You shrugged and said, “if there is something important there, the sender will just send me a letter again”. This is arguably the most liberating advice I've ever been given. Are you still following this e-mail philosophy?

Linus : To some extent, everything remained so, but I changed my workflow so that travel does not interfere with it as much as before. Today I strive to ensure that people do not even notice that I'm on the road. I warn people if I don’t have access to the Internet for more than a couple of days (and this happens in some parts of the world - especially if I’m interested in scuba diving), but most of the time I can work from anywhere in the world. And I try to plan trips in such a way so that they do not overlap with the window when you need to do a merge - then the maximum pull requests fall down on me.

So today my mailbox is stored in the cloud, making it much easier for me to switch between machines, and when I travel with a laptop, I don’t have such a headache as when I needed to download mail to a local computer.

And this is not only with mail - the fact that almost all kernel development is eventually distributed through git means that in principle it’s not so important what machine I work with, and synchronizing work is much easier now than when I worked with patches that came by e-mail separately.

But still, the principle of "if this is important, people will be sent again" is preserved. People know that I work seven days a week and 365 days a year, and that if I didn’t respond to a pull request for a couple of days, it may have been lost in my email chaos, and people send me an email with a reminder to kick me

But now it is much less than before. In 1994, I was not so overwhelmed with work, and when I left for a week, it did not become a problem, but in subsequent years the situation worsened, even to the point that in the old workflow, when patches arrived in the mail, I sometimes I had to miss some of them, because I just didn’t have time, and I knew that people would send them again.

But those times, fortunately, are over. BitKeeper helped me a lot, although not all project participants liked it (and not all used it). Today, thanks to git, I don’t receive thousands of patches via e-mail, and my inbox doesn’t look so bad. Therefore, it is quite simple to keep up with everything.

By the way, perhaps even more important rule than the rule “if this is important, the sender will send the letter”, I consider my long-standing rule: if I don’t have to answer the letter, I don’t answer. If I get a letter, and I think that someone else can do it, I ignore it. Some busy people have automatic answers, where it is written, “sorry, I’ll get to your letter as a result”. And I just ignore everything that, in my opinion, does not concern me. Just because I do not consider it necessary to encourage people to write me more letters.

So I get a lot of letters, but I don’t answer most of them. Speaking really, most of my work is to keep track of what is happening. I see many letters, but usually I don’t write much.

Bob : At a meeting of Linux users in Washington in May 1995, organized by Don Becker [author of Ethernet drivers for Linux / approx. transl.], you stopped in the middle of a performance and asked the audience if anyone knew the results of the Finland-Sweden game at the World Hockey Championship. I, as the representative of Canada, was able to assure you that Finland won. And, by the way, about this: the recent victory of Finland in the junior world championship should have interested you. Or did you support the USA?

Linus : Heh. Hockey is probably Finland’s national sport (and the fact that they played against the Swedes made this game even more personal for me - I speak Swedish because it’s my mother’s native language, but I’m Finn by citizenship), but I’m not a particular sports fan. And the fact that I moved to the United States does not mean that I became interested in baseball and American football, just because of this, hockey has lost its meaning for me compared to the times when my social circle was fascinated by it.

Bob : Many of us admire your desire to call a spade a spade in a public debate about technical solutions for Linux. Others don't like your straightforward style of argument. Do you think you're becoming more or less diplomatic over time?

Linus : If anything, I think I’ve become quieter. I would not call it “more diplomatic”, but I began to better understand myself, and I try to be less assertive.

In particular, because people treat me no longer the same way as before. Previously, the situation was less formal, we were a group of geeks who were having fun and playing. Now the situation is not so. Not such a chamber chamber - now thousands of people are involved in the development, and these are only those who send patches, without taking into account all the people associated with the project.

Part of the change in attitudes towards me is related to the fact that people take me more seriously, which was not the case in 1994. And I don’t complain that they didn’t take me seriously - on the contrary, I grumble about the fact that today they take me too seriously and I can’t freeze any stupid garbage anymore.

So I still say in plain text that some people or companies are doing something nonsense, but now I have to remember that this gets into the news, and that if I show some company a middle finger, this .It doesn't matter if it is worth it or not.

Bob : Do you want to say something more, publicly or in some other way?

Linus : Well, I have never had some kind of “message” that I would like to spread, so ...

Robert Young, and what he has been doing for the past 25 years

Young graduated from the University of Toronto in 1976 with a degree in Historian, and got a job as a typewriter seller. In 1978 he founded his first company, and spent 15 years in Canada at the head of two companies that leased computers. He sold the second company to a larger company, because of which in 1992 he moved to Connecticut to open a representative office in the United States. Shortly thereafter, the new company faced financial difficulties, known as bankruptcy, and Young had to work out of his wife’s wardrobe cabinet.


Robert Young

Although this directly contributed to the founding of Red Hat in 1993, together with Mark Ewing, a young programmer from North Carolina. Young people fell in love with free software, now known as open source software, Ewing, because he could innovate with this software and license to do it, and Young, because he understood how much better clients would work with open technology. compared with closed technologies that were in use at the time. Young served as director of the company, then after entering the stock exchange in 1999 he became chairman of the board of directors, and the director was brilliant Matthew Schulick , and they turned Red Hat into a great business. Now the company is on the S & P 500 list - the index of the largest public companies in the USA