“IT is a rather poor industry” - an interview with Dmitry Nesteruk from JetBrains
Hello. This is again “Without Slides”, and today I have an interview for you with Dmitry Nesteruk aka mezastel , a technical evangelist of JetBrains company. I know Dmitry for quite a long time and, I will not hide, I myself waited a very long time for the moment when we finally go to the studio and sign up. I had a lot of questions for him.
We talked with Dmitry for more than an hour, but did not have time to touch on even half of the topics that I wanted to discuss. What Dmitry managed to tell me:
')
Here is the video, and under the cut, as usual, the text version of this interview.
(at the eleventh minute there is no sound, unfortunately.)
- Dima, I have a question for you - who is a technical evangelist? What are his responsibilities? What is he doing?
- First of all, I must say that we have abandoned the term “technical evangelist”. We switched to another term - Developer Advocate, because evangelism somehow or other, some people have a strong connection in their heads with some religious areas. In general, the “gospel” - “how we carry the gospel” - is too pretentious, I would say. This is the type - here, we have come to save you from bad development practices, therefore - Developer Advocate. That is, a person who upholds certain principles of good design, for example, through, among other things, the toolkit. Therefore, it is a person who is engaged in doing papers at conferences, working on a stand at conferences.
That is, there is a conference - you can put a stand on it and the employees of the companies will show, tell, give all sorts of brochures and interesting gifts, and simply communicate with users. Because one thing is when we communicate with users via the Internet, and another thing is that sometimes you want to get a live feedback. And some people want to see the live developers of JetBrains, to come up and say - "Here, guys, look - I have such a problem, you are not solving it - let's discuss it right here, here, at the conference". And it seems to me great.
In addition, we are doing some marketing materials - in particular, screencasts, webinars ...
- Blogs.
- Blogs, again, yes. That is, on the one hand, we really describe the features of our products, but we are now inclined to do common technological things. If you look at my webinars - for example, the last one - we don’t do such a thing here, “Look, here’s our product and our features, we’ll now decompose what we have there.” That is, we do such things simply as screencasts, and we show the features of the product in the course of some thematic conversation. For example, let's discuss modern C ++, what language features are there - and now our editor is used in the background, some of our features are possible, but we don’t even mention ...
- In general, it is clear that this is a kind of trick, this technique is called product placement and has been known since the time of the first James Bond films. That is, 50 years.
- The nuance is that users, in fact, do not want marketing. Many companies do not understand that if you are doing a conference or event that consists entirely of marketing, then people will have some kind of disgust. Here we do not do that. Therefore, all the reports that I submit at the conference are all just thematic. There is a C ++ conference - I am telling how to do something interesting in C ++. And product placement - yes, it is in the background and it does not bother anyone. That is, if I did something with, say, CLion, but if you want to use Eclipse, then go ahead, no problem.
- Well, okay. How does it feel - evangelism in JetBrains - is it working or not working? Does he help the company or not?
The fact is that we are not the only ones who are engaged in evangelism. Evangelism, for example, is engaged in Microsoft and our approaches are radically different, because Microsoft has a more formal approach - that is, they hire a third-party company, which then deals with the fact that it calls potential users, whether students or business, and asks them: " Guys, we did an event. Do you have an increased desire to install Windows 10? ”.
These metrics are KPIs and these metrics actually make bonuses to evangelists, etc. It is clear that Microsoft has in Russia, because, basically, evangelism, then they have a team more and all this is systematized.
We have a team of six people, on the scale of JetBrains - microscopic. And we really don't have any KPIs. That is, we understand for ourselves what we did is awesome, and what we usually did. That is, if, say, a person has released some awesome thing, which the whole world has noticed - something so cool has done, then we understand that - well, look, this is a achievement. We do not need third-party validation, because sales are sales — that is, we do not have such metrics that “Look, we made an event and after that so many licenses were sold”. We do not think in this way, we just do everything we can.
The same screencasts are needed because you have a product and people want to know what features it has. They can read the dry page or can watch the video. That is, we know that there are screencasts, that there is a need for them, we can watch the viewing statistics. But to understand that these specific screencasts have led directly to sales - it is very difficult. Mathematically very difficult to show and we do not do it. We have an understanding of what it needs, so we do it and try to do it well. And that's all.
- That is, in fact, it is at the level of internal sensation?
- In fact, yes. That is, we see for ourselves when something turns out well. When some thing, relatively speaking, shoots. For example, you do a webinar, and so many people sign up for it that you have to do the second one the same. This is a good criterion. Again, yes, of course, that we, evangelicals, and marketing managers of various teams - we still look at sales, we look at downloading - we keep abreast. In the sense that if somewhere you need to push it, then maybe we will push it. But here is such a direct connection to some metrics that “Look, I sold through evangelism more than you” - there is no such thing.
- Is it good or bad?
- It seems to me that this is good. Because through KPI you can do all sorts of bad things. That is, you can, for example, have an evangelist for a product that is not very commercially successful and, according to KPI, it will have fewer bonuses - because the product is less successful, respectively, and you get less. We don’t have one and, in fact, our products are sold differently - we have very successful ones, we don’t have such successful products. And there are products that evangelicals cover - there are products on which, unfortunately, there are no evangelists yet. As I said, the team is very small, and a lot of products.
- Why the company does not want to expand it?
- The fact is that the search for evangelists is a very difficult process. Because you need a person who: firstly, is technically competent - in the sense that he already has development experience and, in fact, we insist that the person keep writing code. You see, in many companies, without naming names, evangelicals are people who are given slides and they tell them. And that's all.
And we need a person to be technically savvy - to answer any question (that is, if you do, say, a story about C ++, then you should know C ++, you should write on it). And at the same time we want a person to have communication skills, communication skills, so that he would be ready to fly around the world constantly. And something to tell or just chat. And it is difficult. This is a real challenge, because there are not so many of these people.
“So I know Baruch Sadogursky, and he says that you stole the term“ Developer Advocate ”from them. He joked that religion does not allow him to be called an evangelist, and on this occasion he is Developer Advocate. And now this name of your profession has really become a trend. Baruch is a great fellow. In the Java world, he is probably the most famous Advocate.
- You see, this is in some sense show business. Naturally, you go back home and write a bunch of code, debug and dig in all of this, but when you go out to the public - yes, this is show business, how well you can bring some topic.
One more thing that I did not mention, which is also very important to mention, is trips to customers. That is, we, in addition to trips to some events, we also write - either to potential or to existing customers. We say: “Guys, do you want to talk? We will come, we can tell you something, we can sit with a notebook, write down your problems, see how we can solve them better. ”
For example, you come to some hedge fund or investment bank - you need a person who speaks their language. And this is also a problem. Because such people are not found. We have sales, of course, but they will tell you how much it costs, what discounts we can make for you. And you want someone to be alive, who is “Look, I’m cheating here and that’s not working” and we can directly look at this code and do something about it. That is, this is also part of the job.
- If you try to decompose all this according to the percentages of your working time - how do you approximately get it? You probably have this in a team in different ways, right?
- Listen, in fact, everything is not very homogeneous in the sense that there really are events for which we will just go almost every year, and sometimes we take breaks. There are events that year after year are amazing. Suppose you know that there will be 20,000 people for 3 days and therefore it is better not to miss this event. Therefore, there is a certain regularity.
There is still regularity, of course, in releases - that is, when a release occurs, then you must prepare materials for this release. For example, screencasts that describe all the features, then somehow as a bit by bit, some checkpoints, but in addition there are a lot of random things, and the initiative is not punishable.
For example, I sometimes just do webinars - I want to talk about generative art. The way the program generates is really such a small business that my son and I are doing. Generate random images, then framed them with glass and everything. Here, and I'm talking about it. I do a webinar, for example. That is, we have many such spontaneous activities - it seems to me that this is also good, because, on the one hand, it attracts people to JetBrains, on the other hand, there is no rubbing in the product - nothing at all.
- You understand that you ... Just do not beat me and do not poke me with a knife. What are you an absolute geek. That is, a person who is involved in computer art with his son is not to call him another word.
- I will tell you more - I teach my son programming on FPGA, therefore ...
- That is, your son is also a geek!
- I don’t know, at that age it’s too early to talk about geeks.
- This term in general is strange in itself ... Do you have a feeling that you perceive the world - and the industry, and the profession, and the industry - not like the people around you in the same JetBrains, for example?
- It is. And this, I would say, was always, because when I first started working in the IT industry, when I finally decided not to write a dissertation, but to go to work, I ended up in a very strange place. In a place where, let's say, very few people were interested in this The Art of Programming.
- Was it in Petersburg?
- Yes, in St. Petersburg.
- I even know what this place is - we will not call it.
- We will not. And, accordingly ... I would not say that I was more geek then - I just read all the books, I wanted to know everything and, accordingly, I quickly grew, but I quickly got the impression that just nobody was interested in anything around at the level of software, and at the level of hardware, but I was interested in everything at once. That is, accordingly, the feeling that IT, in fact, is not as geeky as it could have been - I still have it. Because we have some specific people who are straight from the technology and want to know everything, including doing things with exotic FPGA, XeonPhi, CUDA, etc. But most people don't need this. Most people just need a salary at the end of the month. And that's all.
- Let this to our listeners and readers seem like a snob-like stance, but on the whole I agree with you. I have the feeling that for many IT it’s just a way of being that way. This is not a hobby and not a hobby ... This is very sad. If we talk more globally - there are a huge number of people in the world who are engaged in what they do not really want to do. I just don’t know how characteristic it is for IT or is it generally for the world?
- In fact, if we talk about our displeasure with life, let's say - I believe that IT is still a normal industry in this regard. For example, we do not have strong physical exertion.
- And in our industry something is paid, after all.
- Yes, something is paid. That is, if you look at the average wages around the world for programmers, this is $ 80,000, that is, you can somehow live on it. In Russia - two and a half times less, but still it can exist. Therefore, it is natural that a private plane is probably not, but for a normal car, an apartment, a summer house, this is quite enough. Therefore, I don’t blame anyone, I’m not saying what’s bad when people go to IT for a steady income.
Another thing is that in the global context, IT is a rather poor industry. Somehow I was visiting the podcast “Debriefing” and honestly said on the air that IT specialist, that is, a typical programmer, is not even the middle class. This is a working class, it is a kind of functional position, because the IT industry itself is by no means a system-building one. It serves other industries - retail, banks, investment companies ...
- Any business.
- Any business. It serves a business in the same way that a business serves an audit, a business serves bookkeeping and all that.
- That is, a typical case is to reduce costs by introducing some kind of automation in some business?
- Yes.
- Of course, I was hooked on the moment now, about the fact that we are a poor industry. If now to some programmer to say that IT is a poor industry, then most likely he will not understand you. Can you reveal this topic a little?
- IT is a poor industry for the reason, for example, that most developers cannot in a good way afford the result of their own labor. We have very interesting technologies, from those technologies that save the soul of the industry, for example, I would call the emergence of SSD, because before SSD work with the disk was just a painful operation if you have some random access, and not just linear operations - everything, death. Since we are doing an IDE, this is a sore subject for us, because a developer without an SSD will simply suffer.
Accordingly, when I make a report on effective development, I talk about how to implement more serious than SSD technologies. And then the director of some outsourcing company comes up to me and says: “You know, we don’t have the money to put the SSD to the developers”.
The most important thing is that at the time when I worked, I felt about the same problems, i.e. you work as a developer, and the company cannot supply you with a normal comp. The company puts the computer, where 4 gigabytes of RAM. Or you come to work, and they give you one screen. I want, for example, six, or at least three. And they look at me like an idiot. And I look at this industry here, as on an idiotic industry in the opposite direction, because I understand that the second monitor is not such an expensive thing.
- I once got a job at Oracle. Great company, but I was also given one 19-inch monitor. But I just took and brought another one of mine, bought it.
- Developers do that.
- Yes, the problem is that I think I was the only one. Those. all the others were completely satisfied.
- Yes. The mentality of the people he is already set up so that that's what our uncle gave us ...
- No, I was not the only one. There was another dude who was my mentor, he is also the right dude in fact.
- I saw people replacing drives with SSDs.
- Or the memory delivered.
- Now, let's say, there are no such shoals in JetBrains, but if you go beyond the limits of JetBrains or a few other good companies and go to outsourcing ... I have been in places quite terrible. It did not work, and been. There, the people are sitting in a sprawling factory, they all have tubular monitors ...
- When it was?
- It was the year in 2008, maybe in 2010 ... But y'know, SSD and monitors are the tip of the iceberg. We are talking about SSD, we are not talking about PCIe-based SSD-carriers, we are not talking about FusionIO-carriers, we are not talking about all sorts of exotic ...
- No, well, PCIe and M2 - are already going into the industry.
- If you go to any company now and ask who is using M2 ...
- Yes, who even uses SSD? I think that already at this stage we will finish our "research".
- In fact, the things that I do, that is, for example, using the GPU, Xeon Phi, FPGA are things that have been around for a long time, you can go to the market and buy them. But the IT industry, it either pretends that they are not relevant, that they do not need to be considered at all, or she says that it is expensive. She just says, well, look, the developer is paying two thousand dollars a month. And you offer for two thousand or five thousand dollars to buy some kind of device, which, like, will accelerate our life, but we don’t really understand how it will allow us to earn more.
- Probably, I absolutely do not understand the industry. Typical margin for Russian outsourcers is 30%. And this is a very highly competitive business. And the truth is that if you start investing in monitors and SSDs, then your margin will fall and you lose the competition.
- I also think that this is just a bad business. Here it is necessary to emphasize such a thing that, in theory, outsourcing is important for the cost of engineering hours. And if they have equipment that reduces the number of these engineering hours, it is not clear what it gives them. Yes, they release faster, the product is possible. But they will get less money.
- They will make this product faster and, therefore, they will have to look for a new client.
- This is one example. Another example: there is a monopolist company. She does not care, the development is two years or twenty years. Because her device will still buy. Accordingly, what do you think, what is her incentive to do something all the time?
- No.
- Zero, absolutely. Therefore, I have such claims to IT. Because IT is an industry that is cunning. She says “look, we're improving.”
.
- “The newest nanotechnology.”
- Yes, but on the other hand, the chief may say to you: “Listen, of course, you showed me a cool thing, but, damn it, let it be somehow without you.” And these are not some fictional stories, these are real stories. Companies, for example, do not want to buy a high-quality IDE, because they are expensive. Although in fact, if you look at the price tag in the context of the same wages ...
But you rightly said about the margin, the margin of 30% is, you know, a good margin. This is more likely the margin of the outsourcer, who works “black”, he is paid in dollars ... More realistic margin for some large “white” outsourcer - 10-15%. And then the investment in IDE will go out of the margin ...
- I think it would never work in Russia. Because if they say to you, they say, dude, you will do business now with a margin of 10% - he will say “yes, go, you nafig!”. We have higher inflation ...
- Perhaps in a large, mass outsourcing 20% is an acceptable amount. Ie yes, of course, people want to make a big profit for themselves and this is understandable right away
- I just don’t understand, maybe you can explain to me ... It seems to me that this business is very complicated. I mean outsourcing, or what is now called Service Company. For example, there is a large company Luxoft. They are pretty cool, they learned quite interesting things to do effectively. Agile-processes have long been lined up with them just the best ...
- Otherwise no one will order them.
- Yes. In this sense, they are wildly steep and in this place they are the engine of the industry. I have a feeling that movement occurs when there are some restrictions that are imposed on you. In this case, the labor market, the market for orders and anything, everything starts to develop. But, unfortunately, they develop a history connected with the processes with the effectiveness of communication only due to the pressure of the market, and at the same time the technological component just sinks.
- And this is a gigantic problem, because people see that their income depends on other things, and this is paradoxical. And very sad. We really are in such a situation, when the latest technologies are needed only for market efficiency, only for the percentage that actually uses them for something. Ie buy some Fusion IO, and throw out all their servers, leaving only a dozen. . .
, , . , , - , - , — , , . , , — . — . , , … ? , , - . , .
— , , , Enterprise-.
— Enterprise- — . IBM, . . .
, Enterprise Enterpris- , - , . Enterprise. Enterprise - , . , , , . - . , .
, , , , - — , , . Rational ClearCase, Rational ClearQuest, IBM- . . . , , . , . : « ? ? - ?».
— ? - ?
— , - . , - , , … , IT, , .
Beats, Apple , — . , , . , , !
— . , . , , , . - JetBrains? - ? , , ? : - , , ?
— , . , . ? — , , . , .
— , JetBrains, , «» . , 19- ReSharper - . «» , . - ? ? !
— . , . ReSharper, . . , , GC, , .
, , , , : « GPU?» , , , , , GPU. Adobe - , CUD. - , , -, .
x86 — . Java- .NET-… x86 , …
— . , Java . - — JIT- .
— . , — — . , , , , , . -
, , , , , SIMD . SIMD, , CPU, — . , , , . compatible , AVX, AVX – , , .
— , Runtime. Runtime , , , AVX ? JIT- , Runtime: « ?» .
— , - , . , , .
— .NET ?
— . RuyJIT. , Mono, .NET, Mono.Simd — Explicit API . , JIT, . Explicit , — . , , , , .
— , Mono API, Mono Runtime ?
- Yes.
— , , .
— : , …
— , , ? . , , - , , , , API, … .
— , , , . , , — OpenMP, . - , , , — , . .
— , , .
— . , SIMD, , . , SIMD . , , , . , , SIMD add, , .
— , , . . , , .
— , . . , , , —. , input output. . - , , , , , …
, , - , x86 , , SSE AVX-. , , - …
— , , .
— , . , - , — . , .
— , : - , , – .
— , , . , GPU, , . , , . , , - . , – - . - , , GPU. , , GPU X86 , - . , .
, , GPU- . , GPU-accelerated, Adobe Media Encoder, Creative Suite, , , GPU - . , - .
— . ZIP.
— , , GPU -. , , , , , . , .
— , , legacy? , . . , .
— , , ? …
— , , - , .
— , , 5+1 …
— 300 .
— , . , . , , . MP3 128 – .
, Beats — , , , - . Hi-Fi, Head-Fi . , , . , - , interference, , , BOSE.
, . BOSE – , . , , , .. , . , , . . , , , .
, — , , 3D… R&D… , - , R&D -, . , .
— CUDA NVidia APU AMD . ? - ?
— , . , . GPU , , . , - , Intel.
— Intel API ?
— . , CUDA. . Intel — Altera FPGA.
— , Intel : , : , . , . , - , — , , . , - . .
— , , GPU, - FPGA, FPGA, , , . , - , . , TP/IP - . , .
, , . , . , - . , . , . FPGA , , .
— , , , - ?
— FPGA — .
— -.
— , . — , . . : «, , . 25 - FPGA. , ». — PhysX. PhysX — NVidia, -, , , AMD- , , NVidia. Windows , NVidia — PhysX AMD. , .
— , 10 .
— , 5-6.
— ? , Half-Life…
— , , Windows7 Windows Vista. , .
, . Xeon Phi , — , , , IT- enterprise?
— , , . Intel. , Xeon Phi . .
— , Xeon Phi … 60 80, , , .
— — Pentium 4.
- Yes. 512 , , , . , , — 256 . , Xeon Phi «», , - Tesla. : GPU, . , . , .
Xeon Phi : , , . Xeon Phi , - x86. , Xeon Phi, , - . , .
, Intel , « 80- . …» . . - 4 . , - , . Xeon Phi, , — - . - - back-end performance — Xeon Phi . Xeon Phi , , .
— — .
— , . Tesla Xeon Phi, .
— ?
— , Tesla .
— Xeon Phi — ?
— , Intel , . , , . , Intel — . , Intel , — . Lenovo , , , , Intel e.
— ? .
— , . - .
— .
— , . , - . AMD , . .
— : 5 «-». Intel , , .
— , Intel . — ! - . , — Haswell, . .
— , , , . !
— , , , , .
- Maybe we just zazhralis? Accustomed to this? We do nothing, and our programs start to work faster?
- This free lunch is long over. The idea of not needing to be optimized for multicore because we will wait, and new processors will come out, it is long over, and a new multi-core era has appeared.
- Everyone said that free lunch is over, everyone began to understand this for a long time, in the class of server-based solutions multicore existed for a long time. The first dual-core AMD came out in 2004-2005, then they were trendsetters. It seems like everyone started to say: "Let's master the multithreading." And the feeling is that by now more or less learned.
- Learned. But we still do not get such a giant benefit, because if the standard is still 4 cores, then the best is that we get this acceleration 3-4 times.
- I understand that the whole history of energy efficiency was needed to learn how to increase the number of cores. But it seems that it did not fire.
“The puncture is that our ecosystem is not scalable.” You can stick one processor on which some cores are in your board. You cannot plug another processor into the board, because there is simply no room.
- In the server - you can.
- Yes, the second processor can be inserted into the server one, but we do not have such linear scaling. Accordingly, if you want to increase power in the context of one machine, you will use Xeon Phi, because this is your only step - to stick a computer into a computer.
I would also note the fact that people more or less understood that it is impossible to increase the requirements for iron to infinity. Look at the games. There was a time when a new game required new hardware. Now it is not so: now you can play the latest StarCraft on hardware five years ago.
- HD appeared, finally high quality.
- There was a time when the new software was living off your exponent resources. Now this less happens. Now there are solutions like Visual Studio, which are still trying to eat gigahertz and gigabytes. But Visual Studio also calmed down, she realized that it’s impossible to devour resources from people endlessly. Thank God, Visual Studio is a 32-bit process, it is limited in this regard.
- How will ReSharper survive in this reality? After all, client projects are becoming more complex, and 32 bits - this, including the memory limit. What to do? Out-of-Process?
- This is essentially the only option. You can't do anything else.
- Does Microsoft have plans to rewrite a 64-bit studio?
- They have a Visual Studio Code, such as we are now starting to write a new IDE from scratch and see what happens there.
- And she is 64 bits right away?
- It seems yes. It would be strange if she were not 64 bit, because it would have been a shot in the leg. The problem is that with large companies like Microsoft or Intel, we never know what their plans are, we don’t know where they want to go.
Here is a Visual Studio Code. Do we want to do ReSharper for him? It is not clear, because Visual Studio Code is a separate topic, its strategy is incomprehensible, and most importantly, well, someone would know it at Microsoft. But it seems that even at Microsoft, sometimes products are released inertially, and not in some clear key. Now they have decided to show the world what they are open-source, which are cross-platform.
- I think this is a business decision. They see that everything has gone to Android, iOS and they, simply without having done so, will lose the market.
- The paradox is that these actions and those announcements that were on the recent Connect - their strategy is incomprehensible. They also have their own mobile operating system, Windows Phone, which has a microscopically small market, and they could continue to develop it, but they still climb into this theme with Android and iOS. But climb or not climb is not yet clear: then something comes out for them, then everyone detains. It seems that Microsoft has no directional strategy. Every department has their own parts, but sometimes these are things that we don’t give us any information about where everything goes.
We, as their clients, cannot adapt in turn. We cannot say that tomorrow we port ReSharper to Visual Studio Code, because we are sure that Visual Studio Code will be an awesome solution. Visual Studio Code can kill the same way as Silverlight.
- That is, for you it is a huge risk and you can not do anything about it? Every time this is a game: to invest or not to invest?
- We are addicted to Microsoft. We had such cases that Microsoft in the last minutes does something such that in ReSharper our code just turns red, and this is announced the day before the release. We do not have time to fit. There's nothing to do with this, because Microsoft has its own goals, which they are not particularly willing to share with partners.
- But there are obvious shifts, for example, another model of announcements by Visual Studio. If you used to work as a partner, received private builds, now builds are public, and you can take all of you as community members. So there is still some movement towards?
- There is a movement towards, but motivation is incomprehensible, it is not clear what is actually happening. Microsoft has really turned its face to Open Source, or is it simply that the market forces them and they understand that they are losing ground, why they are moving in all directions? Another question - if they pull it?
We are still developing .NET in terms of Mac OS and Linux support. The question is, will they finish it? Because there is no guarantee that they will reach it to the state when you can transfer some back-end to Linux so that it works identically and with the same performance. If this is possible, then this is fine, but my question here is: “Where is the money for Microsoft?”. Already there is no Linux of its own to drive .NET. What will they make money on?
- Cloud, it seems to me. Everyone is talking about it.
- Now the story is such that you download the Visual Studio Code in order to write on it under ASP.NET and at the end you will want all this in Windows Azure. Yes, good interesting story. But I remind you that now ASP.NET is not at the forefront of those who write the Web.
The question is that we still perceive Microsoft as a kind of business that makes money just like JetBrains and other companies. At least in the context of JetBrains it is clear: "We do well for developers, we get money for it." The story is over, it is very straightforward.
Microsoft is a bit wrong, because everything is based on them, you, users, want an OS to drive something on this OS. Recent gestures, including Windows 10, they are incomprehensible. I upgraded to Windows 10 on one of the computers and I am very unhappy with the degraded User Experience. You used to write the name of the program, it found you, now for some reason it does not find it.
These are all strange moves. There is an impression that Microsoft is not monolithic, unlike Apple. Apple still has a single strategy, and Microsoft has several departments that each make their own. And then something will work out, something will leak out. The .NET platform and the C # language are a major success. C # is IMHO the best general purpose language.
The problem is that we don’t feel Microsoft has an obvious strategy. And it scares, because you don't know what to invest in you, and what not to invest in.
PS: You can talk with Dmitry personally very soon - at the DotNext conference in Moscow. He will be almost all day at the JetBrains booth and will give two presentations: writing bots and ReSharper 10 .
We talked with Dmitry for more than an hour, but did not have time to touch on even half of the topics that I wanted to discuss. What Dmitry managed to tell me:
- How technical evangelist differs from Developer Advocate
- How does evangelism work in JetBrains
- On the perception of the world by real geeks
- Why IT is a poor industry
- About margins and investments of Russian outsourcers
- What kind of enterprise solutions disgrace the industry and why
- How far is JetBrains from its simple consumer
- How the video game industry was able to "grow" a whole sector of iron
- What is happening now in the market of video cards
- What else can companies come up with to make programs run faster?
- How ReSharper will survive and about Microsoft's 64-bit IDE
')
Here is the video, and under the cut, as usual, the text version of this interview.
(at the eleventh minute there is no sound, unfortunately.)
Technology Evangelist vs. Developer Advocate
- Dima, I have a question for you - who is a technical evangelist? What are his responsibilities? What is he doing?
- First of all, I must say that we have abandoned the term “technical evangelist”. We switched to another term - Developer Advocate, because evangelism somehow or other, some people have a strong connection in their heads with some religious areas. In general, the “gospel” - “how we carry the gospel” - is too pretentious, I would say. This is the type - here, we have come to save you from bad development practices, therefore - Developer Advocate. That is, a person who upholds certain principles of good design, for example, through, among other things, the toolkit. Therefore, it is a person who is engaged in doing papers at conferences, working on a stand at conferences.
That is, there is a conference - you can put a stand on it and the employees of the companies will show, tell, give all sorts of brochures and interesting gifts, and simply communicate with users. Because one thing is when we communicate with users via the Internet, and another thing is that sometimes you want to get a live feedback. And some people want to see the live developers of JetBrains, to come up and say - "Here, guys, look - I have such a problem, you are not solving it - let's discuss it right here, here, at the conference". And it seems to me great.
In addition, we are doing some marketing materials - in particular, screencasts, webinars ...
- Blogs.
- Blogs, again, yes. That is, on the one hand, we really describe the features of our products, but we are now inclined to do common technological things. If you look at my webinars - for example, the last one - we don’t do such a thing here, “Look, here’s our product and our features, we’ll now decompose what we have there.” That is, we do such things simply as screencasts, and we show the features of the product in the course of some thematic conversation. For example, let's discuss modern C ++, what language features are there - and now our editor is used in the background, some of our features are possible, but we don’t even mention ...
- In general, it is clear that this is a kind of trick, this technique is called product placement and has been known since the time of the first James Bond films. That is, 50 years.
- The nuance is that users, in fact, do not want marketing. Many companies do not understand that if you are doing a conference or event that consists entirely of marketing, then people will have some kind of disgust. Here we do not do that. Therefore, all the reports that I submit at the conference are all just thematic. There is a C ++ conference - I am telling how to do something interesting in C ++. And product placement - yes, it is in the background and it does not bother anyone. That is, if I did something with, say, CLion, but if you want to use Eclipse, then go ahead, no problem.
How does evangelism work in JetBrains
- Well, okay. How does it feel - evangelism in JetBrains - is it working or not working? Does he help the company or not?
The fact is that we are not the only ones who are engaged in evangelism. Evangelism, for example, is engaged in Microsoft and our approaches are radically different, because Microsoft has a more formal approach - that is, they hire a third-party company, which then deals with the fact that it calls potential users, whether students or business, and asks them: " Guys, we did an event. Do you have an increased desire to install Windows 10? ”.
These metrics are KPIs and these metrics actually make bonuses to evangelists, etc. It is clear that Microsoft has in Russia, because, basically, evangelism, then they have a team more and all this is systematized.
We have a team of six people, on the scale of JetBrains - microscopic. And we really don't have any KPIs. That is, we understand for ourselves what we did is awesome, and what we usually did. That is, if, say, a person has released some awesome thing, which the whole world has noticed - something so cool has done, then we understand that - well, look, this is a achievement. We do not need third-party validation, because sales are sales — that is, we do not have such metrics that “Look, we made an event and after that so many licenses were sold”. We do not think in this way, we just do everything we can.
The same screencasts are needed because you have a product and people want to know what features it has. They can read the dry page or can watch the video. That is, we know that there are screencasts, that there is a need for them, we can watch the viewing statistics. But to understand that these specific screencasts have led directly to sales - it is very difficult. Mathematically very difficult to show and we do not do it. We have an understanding of what it needs, so we do it and try to do it well. And that's all.
- That is, in fact, it is at the level of internal sensation?
- In fact, yes. That is, we see for ourselves when something turns out well. When some thing, relatively speaking, shoots. For example, you do a webinar, and so many people sign up for it that you have to do the second one the same. This is a good criterion. Again, yes, of course, that we, evangelicals, and marketing managers of various teams - we still look at sales, we look at downloading - we keep abreast. In the sense that if somewhere you need to push it, then maybe we will push it. But here is such a direct connection to some metrics that “Look, I sold through evangelism more than you” - there is no such thing.
- Is it good or bad?
- It seems to me that this is good. Because through KPI you can do all sorts of bad things. That is, you can, for example, have an evangelist for a product that is not very commercially successful and, according to KPI, it will have fewer bonuses - because the product is less successful, respectively, and you get less. We don’t have one and, in fact, our products are sold differently - we have very successful ones, we don’t have such successful products. And there are products that evangelicals cover - there are products on which, unfortunately, there are no evangelists yet. As I said, the team is very small, and a lot of products.
- Why the company does not want to expand it?
- The fact is that the search for evangelists is a very difficult process. Because you need a person who: firstly, is technically competent - in the sense that he already has development experience and, in fact, we insist that the person keep writing code. You see, in many companies, without naming names, evangelicals are people who are given slides and they tell them. And that's all.
And we need a person to be technically savvy - to answer any question (that is, if you do, say, a story about C ++, then you should know C ++, you should write on it). And at the same time we want a person to have communication skills, communication skills, so that he would be ready to fly around the world constantly. And something to tell or just chat. And it is difficult. This is a real challenge, because there are not so many of these people.
“So I know Baruch Sadogursky, and he says that you stole the term“ Developer Advocate ”from them. He joked that religion does not allow him to be called an evangelist, and on this occasion he is Developer Advocate. And now this name of your profession has really become a trend. Baruch is a great fellow. In the Java world, he is probably the most famous Advocate.
- You see, this is in some sense show business. Naturally, you go back home and write a bunch of code, debug and dig in all of this, but when you go out to the public - yes, this is show business, how well you can bring some topic.
One more thing that I did not mention, which is also very important to mention, is trips to customers. That is, we, in addition to trips to some events, we also write - either to potential or to existing customers. We say: “Guys, do you want to talk? We will come, we can tell you something, we can sit with a notebook, write down your problems, see how we can solve them better. ”
For example, you come to some hedge fund or investment bank - you need a person who speaks their language. And this is also a problem. Because such people are not found. We have sales, of course, but they will tell you how much it costs, what discounts we can make for you. And you want someone to be alive, who is “Look, I’m cheating here and that’s not working” and we can directly look at this code and do something about it. That is, this is also part of the job.
- If you try to decompose all this according to the percentages of your working time - how do you approximately get it? You probably have this in a team in different ways, right?
- Listen, in fact, everything is not very homogeneous in the sense that there really are events for which we will just go almost every year, and sometimes we take breaks. There are events that year after year are amazing. Suppose you know that there will be 20,000 people for 3 days and therefore it is better not to miss this event. Therefore, there is a certain regularity.
There is still regularity, of course, in releases - that is, when a release occurs, then you must prepare materials for this release. For example, screencasts that describe all the features, then somehow as a bit by bit, some checkpoints, but in addition there are a lot of random things, and the initiative is not punishable.
For example, I sometimes just do webinars - I want to talk about generative art. The way the program generates is really such a small business that my son and I are doing. Generate random images, then framed them with glass and everything. Here, and I'm talking about it. I do a webinar, for example. That is, we have many such spontaneous activities - it seems to me that this is also good, because, on the one hand, it attracts people to JetBrains, on the other hand, there is no rubbing in the product - nothing at all.
- You understand that you ... Just do not beat me and do not poke me with a knife. What are you an absolute geek. That is, a person who is involved in computer art with his son is not to call him another word.
- I will tell you more - I teach my son programming on FPGA, therefore ...
- That is, your son is also a geek!
- I don’t know, at that age it’s too early to talk about geeks.
About the perception of the world by real geeks
- This term in general is strange in itself ... Do you have a feeling that you perceive the world - and the industry, and the profession, and the industry - not like the people around you in the same JetBrains, for example?
- It is. And this, I would say, was always, because when I first started working in the IT industry, when I finally decided not to write a dissertation, but to go to work, I ended up in a very strange place. In a place where, let's say, very few people were interested in this The Art of Programming.
- Was it in Petersburg?
- Yes, in St. Petersburg.
- I even know what this place is - we will not call it.
- We will not. And, accordingly ... I would not say that I was more geek then - I just read all the books, I wanted to know everything and, accordingly, I quickly grew, but I quickly got the impression that just nobody was interested in anything around at the level of software, and at the level of hardware, but I was interested in everything at once. That is, accordingly, the feeling that IT, in fact, is not as geeky as it could have been - I still have it. Because we have some specific people who are straight from the technology and want to know everything, including doing things with exotic FPGA, XeonPhi, CUDA, etc. But most people don't need this. Most people just need a salary at the end of the month. And that's all.
- Let this to our listeners and readers seem like a snob-like stance, but on the whole I agree with you. I have the feeling that for many IT it’s just a way of being that way. This is not a hobby and not a hobby ... This is very sad. If we talk more globally - there are a huge number of people in the world who are engaged in what they do not really want to do. I just don’t know how characteristic it is for IT or is it generally for the world?
- In fact, if we talk about our displeasure with life, let's say - I believe that IT is still a normal industry in this regard. For example, we do not have strong physical exertion.
Why IT is a poor industry
- And in our industry something is paid, after all.
- Yes, something is paid. That is, if you look at the average wages around the world for programmers, this is $ 80,000, that is, you can somehow live on it. In Russia - two and a half times less, but still it can exist. Therefore, it is natural that a private plane is probably not, but for a normal car, an apartment, a summer house, this is quite enough. Therefore, I don’t blame anyone, I’m not saying what’s bad when people go to IT for a steady income.
Another thing is that in the global context, IT is a rather poor industry. Somehow I was visiting the podcast “Debriefing” and honestly said on the air that IT specialist, that is, a typical programmer, is not even the middle class. This is a working class, it is a kind of functional position, because the IT industry itself is by no means a system-building one. It serves other industries - retail, banks, investment companies ...
- Any business.
- Any business. It serves a business in the same way that a business serves an audit, a business serves bookkeeping and all that.
- That is, a typical case is to reduce costs by introducing some kind of automation in some business?
- Yes.
- Of course, I was hooked on the moment now, about the fact that we are a poor industry. If now to some programmer to say that IT is a poor industry, then most likely he will not understand you. Can you reveal this topic a little?
- IT is a poor industry for the reason, for example, that most developers cannot in a good way afford the result of their own labor. We have very interesting technologies, from those technologies that save the soul of the industry, for example, I would call the emergence of SSD, because before SSD work with the disk was just a painful operation if you have some random access, and not just linear operations - everything, death. Since we are doing an IDE, this is a sore subject for us, because a developer without an SSD will simply suffer.
Accordingly, when I make a report on effective development, I talk about how to implement more serious than SSD technologies. And then the director of some outsourcing company comes up to me and says: “You know, we don’t have the money to put the SSD to the developers”.
The most important thing is that at the time when I worked, I felt about the same problems, i.e. you work as a developer, and the company cannot supply you with a normal comp. The company puts the computer, where 4 gigabytes of RAM. Or you come to work, and they give you one screen. I want, for example, six, or at least three. And they look at me like an idiot. And I look at this industry here, as on an idiotic industry in the opposite direction, because I understand that the second monitor is not such an expensive thing.
- I once got a job at Oracle. Great company, but I was also given one 19-inch monitor. But I just took and brought another one of mine, bought it.
- Developers do that.
- Yes, the problem is that I think I was the only one. Those. all the others were completely satisfied.
- Yes. The mentality of the people he is already set up so that that's what our uncle gave us ...
- No, I was not the only one. There was another dude who was my mentor, he is also the right dude in fact.
- I saw people replacing drives with SSDs.
- Or the memory delivered.
- Now, let's say, there are no such shoals in JetBrains, but if you go beyond the limits of JetBrains or a few other good companies and go to outsourcing ... I have been in places quite terrible. It did not work, and been. There, the people are sitting in a sprawling factory, they all have tubular monitors ...
- When it was?
- It was the year in 2008, maybe in 2010 ... But y'know, SSD and monitors are the tip of the iceberg. We are talking about SSD, we are not talking about PCIe-based SSD-carriers, we are not talking about FusionIO-carriers, we are not talking about all sorts of exotic ...
- No, well, PCIe and M2 - are already going into the industry.
- If you go to any company now and ask who is using M2 ...
- Yes, who even uses SSD? I think that already at this stage we will finish our "research".
- In fact, the things that I do, that is, for example, using the GPU, Xeon Phi, FPGA are things that have been around for a long time, you can go to the market and buy them. But the IT industry, it either pretends that they are not relevant, that they do not need to be considered at all, or she says that it is expensive. She just says, well, look, the developer is paying two thousand dollars a month. And you offer for two thousand or five thousand dollars to buy some kind of device, which, like, will accelerate our life, but we don’t really understand how it will allow us to earn more.
About margin and investment of outsourcers
- Probably, I absolutely do not understand the industry. Typical margin for Russian outsourcers is 30%. And this is a very highly competitive business. And the truth is that if you start investing in monitors and SSDs, then your margin will fall and you lose the competition.
- I also think that this is just a bad business. Here it is necessary to emphasize such a thing that, in theory, outsourcing is important for the cost of engineering hours. And if they have equipment that reduces the number of these engineering hours, it is not clear what it gives them. Yes, they release faster, the product is possible. But they will get less money.
- They will make this product faster and, therefore, they will have to look for a new client.
- This is one example. Another example: there is a monopolist company. She does not care, the development is two years or twenty years. Because her device will still buy. Accordingly, what do you think, what is her incentive to do something all the time?
- No.
- Zero, absolutely. Therefore, I have such claims to IT. Because IT is an industry that is cunning. She says “look, we're improving.”
.
- “The newest nanotechnology.”
- Yes, but on the other hand, the chief may say to you: “Listen, of course, you showed me a cool thing, but, damn it, let it be somehow without you.” And these are not some fictional stories, these are real stories. Companies, for example, do not want to buy a high-quality IDE, because they are expensive. Although in fact, if you look at the price tag in the context of the same wages ...
But you rightly said about the margin, the margin of 30% is, you know, a good margin. This is more likely the margin of the outsourcer, who works “black”, he is paid in dollars ... More realistic margin for some large “white” outsourcer - 10-15%. And then the investment in IDE will go out of the margin ...
- I think it would never work in Russia. Because if they say to you, they say, dude, you will do business now with a margin of 10% - he will say “yes, go, you nafig!”. We have higher inflation ...
- Perhaps in a large, mass outsourcing 20% is an acceptable amount. Ie yes, of course, people want to make a big profit for themselves and this is understandable right away
- I just don’t understand, maybe you can explain to me ... It seems to me that this business is very complicated. I mean outsourcing, or what is now called Service Company. For example, there is a large company Luxoft. They are pretty cool, they learned quite interesting things to do effectively. Agile-processes have long been lined up with them just the best ...
- Otherwise no one will order them.
- Yes. In this sense, they are wildly steep and in this place they are the engine of the industry. I have a feeling that movement occurs when there are some restrictions that are imposed on you. In this case, the labor market, the market for orders and anything, everything starts to develop. But, unfortunately, they develop a history connected with the processes with the effectiveness of communication only due to the pressure of the market, and at the same time the technological component just sinks.
- And this is a gigantic problem, because people see that their income depends on other things, and this is paradoxical. And very sad. We really are in such a situation, when the latest technologies are needed only for market efficiency, only for the percentage that actually uses them for something. Ie buy some Fusion IO, and throw out all their servers, leaving only a dozen. . .
, , . , , - , - , — , , . , , — . — . , , … ? , , - . , .
, enterprise
— , , , Enterprise-.
— Enterprise- — . IBM, . . .
, Enterprise Enterpris- , - , . Enterprise. Enterprise - , . , , , . - . , .
, , , , - — , , . Rational ClearCase, Rational ClearQuest, IBM- . . . , , . , . : « ? ? - ?».
— ? - ?
— , - . , - , , … , IT, , .
Beats, Apple , — . , , . , , !
JetBrains
— . , . , , , . - JetBrains? - ? , , ? : - , , ?
— , . , . ? — , , . , .
— , JetBrains, , «» . , 19- ReSharper - . «» , . - ? ? !
— . , . ReSharper, . . , , GC, , .
, , , , : « GPU?» , , , , , GPU. Adobe - , CUD. - , , -, .
x86 — . Java- .NET-… x86 , …
— . , Java . - — JIT- .
— . , — — . , , , , , . -
, , , , , SIMD . SIMD, , CPU, — . , , , . compatible , AVX, AVX – , , .
— , Runtime. Runtime , , , AVX ? JIT- , Runtime: « ?» .
— , - , . , , .
— .NET ?
— . RuyJIT. , Mono, .NET, Mono.Simd — Explicit API . , JIT, . Explicit , — . , , , , .
— , Mono API, Mono Runtime ?
- Yes.
— , , .
— : , …
— , , ? . , , - , , , , API, … .
— , , , . , , — OpenMP, . - , , , — , . .
— , , .
— . , SIMD, , . , SIMD . , , , . , , SIMD add, , .
— , , . . , , .
— , . . , , , —. , input output. . - , , , , , …
, , - , x86 , , SSE AVX-. , , - …
— , , .
— , . , - , — . , .
— , : - , , – .
— , , . , GPU, , . , , . , , - . , – - . - , , GPU. , , GPU X86 , - . , .
, , GPU- . , GPU-accelerated, Adobe Media Encoder, Creative Suite, , , GPU - . , - .
— . ZIP.
— , , GPU -. , , , , , . , .
, «»
— , , legacy? , . . , .
— , , ? …
— , , - , .
— , , 5+1 …
— 300 .
— , . , . , , . MP3 128 – .
, Beats — , , , - . Hi-Fi, Head-Fi . , , . , - , interference, , , BOSE.
, . BOSE – , . , , , .. , . , , . . , , , .
, — , , 3D… R&D… , - , R&D -, . , .
— CUDA NVidia APU AMD . ? - ?
— , . , . GPU , , . , - , Intel.
— Intel API ?
— . , CUDA. . Intel — Altera FPGA.
— , Intel : , : , . , . , - , — , , . , - . .
— , , GPU, - FPGA, FPGA, , , . , - , . , TP/IP - . , .
, , . , . , - . , . , . FPGA , , .
— , , , - ?
— FPGA — .
— -.
— , . — , . . : «, , . 25 - FPGA. , ». — PhysX. PhysX — NVidia, -, , , AMD- , , NVidia. Windows , NVidia — PhysX AMD. , .
— , 10 .
— , 5-6.
— ? , Half-Life…
— , , Windows7 Windows Vista. , .
, . Xeon Phi , — , , , IT- enterprise?
— , , . Intel. , Xeon Phi . .
— , Xeon Phi … 60 80, , , .
— — Pentium 4.
- Yes. 512 , , , . , , — 256 . , Xeon Phi «», , - Tesla. : GPU, . , . , .
Xeon Phi : , , . Xeon Phi , - x86. , Xeon Phi, , - . , .
, Intel , « 80- . …» . . - 4 . , - , . Xeon Phi, , — - . - - back-end performance — Xeon Phi . Xeon Phi , , .
— — .
— , . Tesla Xeon Phi, .
— ?
— , Tesla .
— Xeon Phi — ?
— , Intel , . , , . , Intel — . , Intel , — . Lenovo , , , , Intel e.
— ? .
— , . - .
— .
— , . , - . AMD , . .
— : 5 «-». Intel , , .
— , Intel . — ! - . , — Haswell, . .
— , , , . !
— , , , , .
, ,
- Maybe we just zazhralis? Accustomed to this? We do nothing, and our programs start to work faster?
- This free lunch is long over. The idea of not needing to be optimized for multicore because we will wait, and new processors will come out, it is long over, and a new multi-core era has appeared.
- Everyone said that free lunch is over, everyone began to understand this for a long time, in the class of server-based solutions multicore existed for a long time. The first dual-core AMD came out in 2004-2005, then they were trendsetters. It seems like everyone started to say: "Let's master the multithreading." And the feeling is that by now more or less learned.
- Learned. But we still do not get such a giant benefit, because if the standard is still 4 cores, then the best is that we get this acceleration 3-4 times.
- I understand that the whole history of energy efficiency was needed to learn how to increase the number of cores. But it seems that it did not fire.
“The puncture is that our ecosystem is not scalable.” You can stick one processor on which some cores are in your board. You cannot plug another processor into the board, because there is simply no room.
- In the server - you can.
- Yes, the second processor can be inserted into the server one, but we do not have such linear scaling. Accordingly, if you want to increase power in the context of one machine, you will use Xeon Phi, because this is your only step - to stick a computer into a computer.
I would also note the fact that people more or less understood that it is impossible to increase the requirements for iron to infinity. Look at the games. There was a time when a new game required new hardware. Now it is not so: now you can play the latest StarCraft on hardware five years ago.
- HD appeared, finally high quality.
- There was a time when the new software was living off your exponent resources. Now this less happens. Now there are solutions like Visual Studio, which are still trying to eat gigahertz and gigabytes. But Visual Studio also calmed down, she realized that it’s impossible to devour resources from people endlessly. Thank God, Visual Studio is a 32-bit process, it is limited in this regard.
How ReSharper will survive and about Visual Studio 64 bit
- How will ReSharper survive in this reality? After all, client projects are becoming more complex, and 32 bits - this, including the memory limit. What to do? Out-of-Process?
- This is essentially the only option. You can't do anything else.
- Does Microsoft have plans to rewrite a 64-bit studio?
- They have a Visual Studio Code, such as we are now starting to write a new IDE from scratch and see what happens there.
- And she is 64 bits right away?
- It seems yes. It would be strange if she were not 64 bit, because it would have been a shot in the leg. The problem is that with large companies like Microsoft or Intel, we never know what their plans are, we don’t know where they want to go.
Here is a Visual Studio Code. Do we want to do ReSharper for him? It is not clear, because Visual Studio Code is a separate topic, its strategy is incomprehensible, and most importantly, well, someone would know it at Microsoft. But it seems that even at Microsoft, sometimes products are released inertially, and not in some clear key. Now they have decided to show the world what they are open-source, which are cross-platform.
- I think this is a business decision. They see that everything has gone to Android, iOS and they, simply without having done so, will lose the market.
- The paradox is that these actions and those announcements that were on the recent Connect - their strategy is incomprehensible. They also have their own mobile operating system, Windows Phone, which has a microscopically small market, and they could continue to develop it, but they still climb into this theme with Android and iOS. But climb or not climb is not yet clear: then something comes out for them, then everyone detains. It seems that Microsoft has no directional strategy. Every department has their own parts, but sometimes these are things that we don’t give us any information about where everything goes.
We, as their clients, cannot adapt in turn. We cannot say that tomorrow we port ReSharper to Visual Studio Code, because we are sure that Visual Studio Code will be an awesome solution. Visual Studio Code can kill the same way as Silverlight.
- That is, for you it is a huge risk and you can not do anything about it? Every time this is a game: to invest or not to invest?
- We are addicted to Microsoft. We had such cases that Microsoft in the last minutes does something such that in ReSharper our code just turns red, and this is announced the day before the release. We do not have time to fit. There's nothing to do with this, because Microsoft has its own goals, which they are not particularly willing to share with partners.
- But there are obvious shifts, for example, another model of announcements by Visual Studio. If you used to work as a partner, received private builds, now builds are public, and you can take all of you as community members. So there is still some movement towards?
- There is a movement towards, but motivation is incomprehensible, it is not clear what is actually happening. Microsoft has really turned its face to Open Source, or is it simply that the market forces them and they understand that they are losing ground, why they are moving in all directions? Another question - if they pull it?
We are still developing .NET in terms of Mac OS and Linux support. The question is, will they finish it? Because there is no guarantee that they will reach it to the state when you can transfer some back-end to Linux so that it works identically and with the same performance. If this is possible, then this is fine, but my question here is: “Where is the money for Microsoft?”. Already there is no Linux of its own to drive .NET. What will they make money on?
- Cloud, it seems to me. Everyone is talking about it.
- Now the story is such that you download the Visual Studio Code in order to write on it under ASP.NET and at the end you will want all this in Windows Azure. Yes, good interesting story. But I remind you that now ASP.NET is not at the forefront of those who write the Web.
The question is that we still perceive Microsoft as a kind of business that makes money just like JetBrains and other companies. At least in the context of JetBrains it is clear: "We do well for developers, we get money for it." The story is over, it is very straightforward.
Microsoft is a bit wrong, because everything is based on them, you, users, want an OS to drive something on this OS. Recent gestures, including Windows 10, they are incomprehensible. I upgraded to Windows 10 on one of the computers and I am very unhappy with the degraded User Experience. You used to write the name of the program, it found you, now for some reason it does not find it.
These are all strange moves. There is an impression that Microsoft is not monolithic, unlike Apple. Apple still has a single strategy, and Microsoft has several departments that each make their own. And then something will work out, something will leak out. The .NET platform and the C # language are a major success. C # is IMHO the best general purpose language.
The problem is that we don’t feel Microsoft has an obvious strategy. And it scares, because you don't know what to invest in you, and what not to invest in.
PS: You can talk with Dmitry personally very soon - at the DotNext conference in Moscow. He will be almost all day at the JetBrains booth and will give two presentations: writing bots and ReSharper 10 .
Source: https://habr.com/ru/post/271887/
All Articles