Ready solutions # 2. Alexey Lustin. The path of "Zen" from Java to 1C
Away - Alexey Lustin
Presenter - Sergey Gorshenin
Dear friends, I welcome you to the “Ready Solutions” program. I represent our guest - this is a person with twenty years of experience in the field of information technology, IT specialist, technical director and just a good person. Our today's guest is Alexey Lustin.
Hello, Sergey. Hello colleagues.
______________________________________________________________________________________________________________________
Chronological reference
')
Aleksey Aleksandrovich Lustin was born in Anadyr in 1980 and studied at the Voronezh Technical University with a degree in CAD Engineering. Since 1995 - IT-specialist. He started out as a 1C programmer in the city of Lipetsk, was engaged in the automation of business processes in the Yuterra network. Since 2008 - leading specialist Java, Rails. In 2008 he moved to Moscow. From 2012 to 2014 - Deputy Head of the Department for Innovative Development and Systems Analysis at the Svyaznoy Group of Companies. Currently he is a technical director, trainer, architect-consultant of the Silver Bullet company.
Alexey, for a start, I suggest you to tell a little about yourself in order to designate what we will talk about in the future.
It is quite difficult, like any beginning. In short, I speak about myself in such a way that I was born in Chukotka, far, far away, near the Arctic Circle. IT has been doing it since 12 years, that is, in fact, not 20, but in my opinion, already 22 years old. He has gone from the usual refilling of cartridges, the operator for filling out incoming invoices, to the architect, technical director and so on.
How did I go through it? Here, probably, it is necessary to tell longer.
I came from Chukotka in 1997 to go to college, because the main group of good teachers was concentrated in the central region, and not in the Far East. Maybe even in Novosibirsk, but in general, at that time, high-quality IT education could be obtained in technical cities - either in Moscow, or in Voronezh, or, of course, in St. Petersburg. Unfortunately, this was not the case in the Far East. To provide myself with money, and the time was hard, I took on any work on IT. This is the administrator of a computer club, and a collector of computer equipment, and the developer of 1C, oddly enough, even in the old-old version. I also tried my hand as a Java developer, when it was just beginning to develop, and as a Clipper database developer, who still remembers such things. For those times it was fun enough.
And after the end of the student phase, serious life had already begun, and there I went a different way: I was both a programmer and a project manager, depending on what the global goal of the employers I worked for (business owners, IT directors or IT executives in general).
Well, and then, in Lipetsk, I worked for the company Yuterra - this is a fairly large Russian holding company that combines a network of non-food goods stores - at that time there were about 112 supermarkets from Blagoveshchensk to Kaliningrad.
A little later, in order of my own development, I moved to Moscow on my own initiative, closer to the grandees of our industry, in order to really gain experience and share my approaches. So I got to Svyaznoy and worked there for about 4 years. I was there both a leading programmer and an internal technical leader; last year my role in the company was called “1C Competence Center”. So you understand, there are about 200 specialists in 1C.
Well, knowing the Messenger, I understand the level, yes.
Yes. Here is a way of life. In terms of my personal relationship with him.
So, in the 95th year you started as an IT specialist. Do you think that the model of an IT specialist of the 95th year differs from the IT specialist of 2015?
In general, nothing has changed, in my opinion.
But what about the changes in paradigms, in concepts, in technology development?
Paradigms have not changed. All the paradigms on which we are now parasitic, invented by the four authors - this is Brooks, Wirth, Ward and Hyudak. In fact, they invented these paradigms in 93-94 years (I am not talking about software engineering, which was invented in the 70s). Everything that is happening with technology is laid there. Nothing changes. On the other hand, only the practice of applying these paradigms is growing, that is, an IT specialist who started in 95th year has, conditionally, 20 years of experience in applying these paradigms. This is his only competitive advantage over a person starting in 2005. Just a person starting in 2005, has 10 years of practice. That's the whole difference. And the paradigms are the same.
Good. And let's remember the 95th year. Still, IT people then - these are people who loved IT, they came to this field at 12, at 15 years ... It is no secret that now some of IT people choose IT only because it is fashionable.
Yes, now really big fashion for IT. But then they chose, because it was fashionable. For example, you can recall the “dotcom boom” of 2000, which affected Russia, because even then there was a fashion for IT. Yes, and 3 years before, in '97, the fashion for engineering specialties was also very large.
I will explain why. In Russia at that time there was a general lack of IT specialists of any nature. For a good IT-specialist were willing to pay big bucks. They were no longer willing to pay for lawyers, economists, but they were ready for IT. It was quite an interesting time. If we recall that in the 93rd year, the passing score for engineering specialties was 9 - three triples, then in the 97th year for engineering specialties related to IT, the passing score was already around 14-15 - one four, two fives. This is an indicator. And, in my opinion, this passing score is still not reduced. That at the beginning of the two thousandth, that now the passing score and requirements for IT-specialties remain at the level of 97, 98, 99-ies.
On the other hand, and then in IT they went, including for fashion. Therefore, now many of those who started back then are no longer IT specialists. Some of them are businessmen, someone became an economist, someone left.
There were not those who came for fashion, but those who really loved it. And if we see an IT specialist with ten years of experience, then he most likely went through all the crises of this growth - 1, 3, 5 years. As in family life, you married IT, then a year later you realized who you married, after 3 years you had a falling out, and after 5 years you finally reached a certain “Zen” when you understand that you are from each other need to.
Such a philosophy.
You have a rather interesting experience, but you cannot say that you are an IT specialist, although you yourself insist on it. Why?
I have two theses that reflect my experience and the general concept of what I am for a specialist: this is Overskilled — the so-called “skill overkill” (there are too many), and at the same time Troubleshooting is a “person - problem solver”. I usually say that I can solve any problem in IT. Due to the experience, due to the technique, due to the developed algorithms - in general any. Because any problem can be solved. It doesn't matter if this problem is related to 1C or not. There can be no unsolved problems. This is the first.
Since I live with such an approach for a very long time, can you imagine how many problems I have already solved? That is why, in fact, I am overkill on skills. I can solve the problem with Ruby, with C ++, with the Assembler I can even, if there are problems. With 1C, of course, with business applications, with applications written by third-party developers, with OpenSource, Commerce applications. I have just experienced it all. Cartridge accounting? Yes please.
Then I have a counter question. My career began in the distant two thousandth, and I started exactly with 1C. But after 6-7 years, I stopped being a single member in the truest sense of the word, although, probably, somewhere deep down in my soul, I still program at the seven. But why are you with such a rich technical experience still consider yourself to be one?
In fact, I usually call myself a business application specialist. And in Russia, and not only in Russia, it is impossible to create business applications without the 1C platform.
Can we say that in this case 1C is the most common platform for creating business applications, at least in Russia?
1C is an implementation of the concepts of a problem-oriented language. Let's go back to the 95th. There was a problem. There are a lot of businesses. Imagine how many configurations are now in Russia, written from scratch, without taking into account the typical. That is, de facto, we can say that every 1C specialist wrote his configuration from scratch. Bad, good is not important. On the INFOSTART.RU portal, 450,000 active users, 1C specialists are now registered, that is, we can say that each of them has written his own configuration.
You know, I will not go far, for example, I have 4 full-fledged software complexes. This is not counting the minor developments and reports.
I am usually a supporter of saying that how many 1C specialists are, so many atypical 1C configurations written from scratch. And how many businesses in Russia? They are much more.
Probably all the same business models?
Yes, business models. Most often, the business architecture is directly proportional to the architecture of the team creating this business. There is no the same business and the same process also does not happen. Even if they are based on the same principles, the concrete implementation is always different. We must accept this.
Where did the 1C platform and its language for developing business applications come from? All this appeared in response to the problem that the business grew in Russia - there was a transition from the old system to the new: many small enterprises appeared, the growth was very rapid. There were many small businesses. They could not all work on "template" business practices ("bestpratiksam").
Then there were no such practices as such in Russia.
Yes, there was no consulting as such for this market yet. But it was necessary to somehow work.
There is a need for a language to quickly describe business processes. Because you will not find in any platform a language in which you can write “Documents. Entry Bill. Pass ()” (register, print, send to printer, read off sale and so on). And it had to be described.
Conventionally, the idea was that even an accountant would be able to describe his business process. It didn’t quite work out - anyway, special people were needed who had an algorithmic base and could describe business processes in a special DSL language (problem-oriented language). But this way it was possible to reduce costs and quickly automate any business from 5-10 people.
In essence, this saved time and expenses for the creation of these base classes, base models, because they were already in fact.
Yes, it was a huge leap. The fact is that at that time, in the 95th year, the main language for developing business applications was C ++, which had a low level of abstraction.
Then on Clipper a lot was written, as I recall.
Clipper was too low-level language. Then Java just appeared, as a kind of framework that raised the level of abstraction. But he raised this level only up one stage - he allowed the component model to build, and 1C jumped right up through 5 levels, immediately to the level of a businessman. Even now, after 20 years, neither abroad nor in Russia there is still no such language that would allow to describe business processes directly.
Now banks are trying to create a language for automating traders and describing their algorithms in a certain DSL language - they are trying to do this in the Java world (or, more precisely, in the functional world). And we have such a language, conditionally, has existed for 20 years, so we are in a favorable position here. And it is unlikely that a language of the same level will appear there in the near future, making it possible to level up the system’s work at basic levels so effectively that at the first stage it is possible not to dwell on problems with the code, memory, so that you can immediately think about the benefits of business.
Remember, why did you come up with Agile? In order for developers on third-party platforms to start thinking about the benefits for the end user. They didn’t do space memory management algorithms, but thought how these algorithms can be useful for business. Not a businessman in a shirt (the owner, who invests money in it), but the end user - the storekeeper, the operator, the accountant.
There is still such a moment: how to justify for a business the need to optimize some algorithm in order to win one and a half megabytes of memory? In C ++, it is almost impossible to prove the feasibility study of such optimization for a business. And with this it’s easier, his tasks are close to business goals, so he can explain what will become faster, what will improve (and he can implement the optimization from memory just with C ++).
And if you go back to why I consider myself a 1C specialist - I always work with businessmen. Initially, 20 years ago, I started at the Mandatory Medical Insurance Fund. My first task was to automate the collection of payments on CHI. My first customer was the head of the OMS Fund, he explained the strategic business goal: why does he need it and so on. And the technical implementation and the technical risks associated with it were completely on my side - the choice of platform, whether it would be Access, Clipper, and so on, completely fell on me.
This was my first task - at that time I was 15 years old and there were no specialists in our city in databases at all. Besides me and Andrew, my colleague is a good C ++ programmer. And in our team there was a “senior comrade”, a radio technician of 35 years old - he explained to us how to implement everything technically, and then “combed” what was done. And we, the young guys, took the strategic goal of the person who was responsible for all district compulsory medical insurance, and tried to implement it. Bad, good, with errors - it doesn't matter anymore. Probably, if he sees us now, he will say that we all did badly there. But it normal.
Is it possible to say that it was for you that you decided then that you need to develop, that by setting your goal at the head of your work, he gave you an impetus to further development?
Yes, of course, then we were strongly influenced by what he believed in us. We were young tenth-graders, we won at the Olympiads, such IT specialists to the bone marrow - but, in fact, we were boys with no work experience. And he took a chance and gave us the opportunity to try to do something. And we tried, yes. In order to be close to computers and generally be close to everything. It was such a good motivation.
I remember the beginning of my career - it was Vodokanal, there was 1C. I found 1C under DOS version 2.0, then moving to 7.5, to 7.7. But when I started writing something under 7.7, I very quickly realized that I was missing the very accounting background. I then went to a three-month course, and this, for me, probably, on the one hand, served as an impetus for further development, and on the other hand, it served as a kind of click to get me to start moving from programming further into the depth of business processes. Is it possible that IT-specialists who write programs or applications for various business tasks will move to get away from programming and move to adjacent planes, to other platforms, for other tasks or to another area in general?
Yes, it usually happens. The world is divided. At first you just have some programming skills. You can write on Pascal, on C ++ something, some algorithms. This is a basic skill - at this stage you have no experience in working with business problems. But you get 1C-platform. You try to solve business problems with it, pass it through yourself and your first level appears. And what next - you decide. You can go to businessmen - it happens. You might just realize that IT is not yours.
You can go to business methodologists (or else this is called an analyst-methodologist). This is not the one who describes the methodology, but the one who is trying to be a link between the developers (those who create a product) and business customers. This is some kind of analyst who studies business processes, tries to find hidden requirements, and possibly attracts external methodologists.
Is this probably something close to the architect?
No, it is more related to system analysis. The architect - he builds component models, and this specialist is a kind of designer, he is more involved in analyzing exactly how to build the final model.
One way or another, you begin to specialize. You may want to leave the 1C-world at this moment, because suddenly you decide that you can design everything yourself on some low-level software. Take, for example, Java, as it is free, and you start to design something on it. As a result, people either begin to do something else, or, what happens most often, return back to 1C, but with different knowledge.
The amount of experience is directly proportional to the number of mistakes made as they grow. This is such a feature. Therefore, in the end, you still in most cases return to the 1C world. Maybe in a different capacity - for example, as a business customer, or as a chief architect, looking at 1C from the outside, but still working with her. Because 1C is a platform for automating business processes, and you don’t go anywhere on your own.
But what about the reputation of one person? I remember when I started, many of my classmates laughed at me. Say, here we are programming in normal full-fledged languages, and you program in a language that is not enough that is made in C, it is also written in Russian.
Unfortunately or fortunately, I had a lot of teams in my work that are not the same. In general, I sometimes consider myself to be a specialist in Ruby on Rails and Java in addition to the 1C specialist - it so happens that I have experience writing applications on these frameworks. But there is another problem. People who write on .NET, Java, on anything, they are too far from the people. They do not automate the business. They write certain components for business, or rather components for those programmers who will then build a business application from these components.
In fact, they build cubes, from which further someone will build the final product.
Yes, this is how it works. After all, we never think about the fact that there are people who write programs in Assembler that implement the logic of microchips. And there, too, there are internal algorithms - various southern, northern bridges.
Come on easier. Here we control the mouse every day, we do not think about how the mouse movement is transmitted to the screen.
Yes, somewhere there are people who program microcrystals inside microcontrollers. There is a specialty "Physical Chemistry", where they learn to work at the level of microcontrollers, to program there. For them, 2 bytes is a lot. They have a memory there is not gigabytes measured, and bytes. These people give their finished components to other people who already describe working with devices at a higher level, for example, in C ++. Then some other people take these components, written in C ++, and make of them their applications, components, or libraries — already on .NET or on Java, or on something else. And then they come together, take these components and make them the ultimate business application. This is how it works.
Let's go back to the singles. Is it possible to say, based on your words, that a certain contradiction arises, that, on the one hand, you don’t want to be single member, but you have to.
You have to be one person, but you don’t have to be one person. This is the contradiction. For convenience of understanding, you just need to remove the word 1C. 1C is CJSC 1C, it is a company that develops, which invests both money and methodology inside this platform. You have to be a specialist in business automation, but you do not need to be a specialist in business automation. Resembles schizophrenia. But in general, you just have to learn how to switch the toggle switch.
That's when you communicate with a business customer, your task is not to discuss with him any technical issues - technical debts, technical architecture. The customer does not know how to build a cluster horizontally or vertically, how to build racks and so on. He tells you a certain wish. In general, the customer wants to get some of the alleged visual behavior of the mold as a whole. Unless, of course, this is not the customer who orders your compiler.
... Let's just say that if your customer is a business representative, not a representative of the IT industry.
Yes. We are talking about the final business. , . – , - – , , . .
IT-, , , -. . , 1 – , , ( , ) . , , , , .
. 1-. , , ?
. 1- , . , . , - , . 1- .
, , ?
, , . 1 . , , . – , . , , . , , , .
80%, , 2 , 20% . .
. . , , 1: , , « », . 10 , , -, .
1 . . : - , 100 , , . , 1- . 1 . , , .
, ?
Yes. , . . 1- , IT, ( , ) — , -, . .
, , .
. 1- , , -. .NET- , IT, , .
«» , : ? , , , , . ?
-. -, , . , , .
10 . 1-, : , 1, 80% 1 - – Android, Clojure , . , . JavaScript, HTML. L HTML language — .
, 1. , , , XML 8 . -, WSDL (Web Services Description Language), L — . , 1, , , . .
, , , , , 5 .
5 , – - - , . , 6 , , , 20 , , . ( ) .
, , , . , , , Python, .NET, 1 XML, 1 HTML… , , 1 Web. , 1 - . - , 1, - «».
, , 2014- – , , . 1- (, , , , ). , « » — . . «», , .
, PHP , , . , , , , . Node.js Ruby on Rails, . , , - … .
, – «» 1-. , , , ?
– - . 1 «1», , . .NET, Java, Phyton . , 95- , , (, , ).
, . , – . , DSL , . . , : 1 .NET, — . .
. ( ) .NET Java 1- 5 . ( , , ) – 1-, , , .
, , , , . – .
, - ( ) : «, .NET, ». . , , : «, 1, ?» 1- .
: 1, , : « — .NET», .NET- : « 1, 1 — », . .
. , IT-. , , , , , 4 , 4 , , 2, , . , , IT-? , . , , .
, , , , .
IT- : , . IT- , – , IT- , , . . () — .
, ? , , — . , «», . ? , , .
. : , — , - . -, : « , » – , , .
IT- : . .
, , , . - .
. , 1-. ? 1- , ?
. : , , WMS – 7 . «», , Java, Ruby on Rails ++ . , : « (DZBZ2 Godex ), , ».
— 1, , – , ++, Linux. – . , … . , «».
. , . : , . , . , .
, , «», , WMS 1 . — . , , . – Java-, , , Java 1.
, 1 – , 1- - (, ..), .
, 1- Java-. – , .. , «1 – ». , , , -. , , WMS- 1 . , . . , – .
, .
, – , . , .
: , , , . : « ». : « -, , ». : « -? . 2 , . , . ?» , .
, , , -, ( ) .
I really liked how you touched the topic of costs. Recently, perhaps due to cloud computing, a situation has often emerged when the development team is being carried out somewhere in the regions. I, as an IT director, see in this, first of all, a reduction in costs. What are the pros and cons in this you see, precisely from the point of view of the technical director? Can you say that code quality suffers from this? Or is it all about the cost?
Quality suffers from this for sure.
Is the quality of the code unequivocally affected?
Everything suffers. First, when you have a non-distributed team, it’s almost impossible to manage. The only one who can manage this is, in my opinion, Alexander Belov. He has the only successful management experience.
The developers of this team should have a special psycho. You cannot hire a person with the nature of a stationary developer somewhere in Izhevsk or Yekaterinburg. You need to look for a completely different person who will be able to work in a remote team, reporting to the “center”.
In a distributed team, the costs of manageability are generally increasing. It seems that you save on money, on a payroll (payroll) due to the fact that you are going to the regions, but you start to lose controllability.
When you build your team, in any case you need to make a separate small development department methodological, detached, with your manager, with an office manager, and so on. Then it somehow works. Then this cost center pays off because it makes the final product. This is from practice.
If someone tells you that development in a distributed team will be more profitable, then they lie, they simply say that for money, according to payroll it will be less expensive. But the total cost — this is called total cost of ownership — is likely to be higher in remote development. And it's not about quality, but about the costs that you will bear.
That is, I correctly understand that they often say that direct costs - these initial investments (if you can say so) will be less, but about indirect costs that affect common bones - are they either forgotten or specifically kept silent about them?
Usually, when an office outside is taken out, people are aware of the situation with indirect costs, something like this: wow, it turns out, indirect costs are much higher than in Moscow, St. Petersburg or other central cities.
Is there a difference between simple developers and 1C developers from the point of view of remote companies?
From practice I have not seen. Money is different, but the interest is the same.
At the remote office, you still need a “champion”, the cost of which will be two times higher than the typical average salary level. And in the end, you will need to compensate for this small leadership at the expense of office costs, maybe even reduce the costs of marketing, accounting. But sometimes an accountant will need a local just in case, or an office manager. Practice such.
Well I myself come from the regions. I’ve been in Moscow only for the last 4 years, and before that I had shook from Rostov to Pskov and even in Siberia, so it happened that I was for a while. I mainly worked in the regions.
Good. And if you use any modern tools for group development in the clouds, where everyone works in one place and does not communicate live?
First, for 1C of the world, this has not yet been created. There are our attempts in collaboration with the guys, specialists in Linux, now to deploy the cloud configurator - a special system for group development of configuration in the cloud with a customized environment, with all the tools. This configurator is currently undergoing internal beta testing. Maybe it will be possible to roll it out in the near future or the 1C company itself may possibly do something similar in the future.
Do I understand correctly that this is just a kind of virtual machine running the configurator, then the question is how is multiple access organized there, how is the joint development in it going?
So let's separate the flies from cutlets. We remove the word 1C. Infrastructure developer in the cloud - this is a big problem now in the entire market. In 1C, in Java, in .NET, in Java Script and so on.
Now there is such a concept Vagrant. 1C: Bitrix uses it (just started using it). This is when a little virtual tuned machine comes to you with a deployed Bitrix - so that you don’t waste time, but immediately start coding. In the same way, you can deploy Bitrix locally on your virtual machine, or somewhere in the cloud - yours, if you have one, or in a private one - whatever. As a result, we operate with these small virtual operating systems that come preinstalled to you and are automatically updated. Such a trend is everywhere. Therefore, the word 1C is removed - this is everywhere now.
This reduces costs. For example, to buy iron. Because if we imagine some ideal stationary computer of the 1C developer, then it will now cost, probably, about 40 thousand rubles, if you buy it very, very cheap. Although at current prices even, probably, already 60. There should be something like i7 (multi-core), 16 gigs of RAM and a terabyte SSD screw to work with the configurator cache so that everything fits.
Well, yes, software requirements are growing faster.
Yes, and it is necessary to recoup 90 thousand rubles. If you work in Moscow, you can somehow do this, and in the region 90 thousand rubles, even with a loan, is quite difficult to recoup.
And when using a remote environment, you just need a browser or some stationary client that will work with it over the network. With the development of communication channels (and they are still developing), this will be quite good and convenient. Therefore, it is a really interesting trend.
And how in your case with the cloud configurator will the speed of the channel be affected? It seems clear that this will affect the display of the picture, but how good will the performance on the cloud side be? And what will it depend on?
From practice - the performance there will be 3-4 times higher than, say, on your local machine, even with an SSD. Because the performance of the solution itself is already dependent on the supplier of this solution.
All of our final business solutions, final consulting, practices, we still sell. But the tools for people to develop, we lay out in OpenSource.
How was our community formed? We shared our tools, our work. Yes, and Infostart was so formed. All tools are laid out. So we, it turns out, are debugging it. The toolkit (the same Snegopat), is in many ways more productive than 1C, several times. Working with the configuration, working directly with the 1C storage is also implemented there several times faster than regular work. Due to the fact that there are good C-shniki, assemblers sometimes come across.
Well, yes, Mercedes AMG, he, of course, better than just a Mercedes, because it is brought to mind.
Yes, for example, there is an opinion that 1C on virtual machines slows down and reduces the performance by a factor of two. But there is another opinion that if properly configured virtualization, it will not lose in performance. And if you correctly configure the parallelism of the threads, that is, to implement the code of the internal tooling in a slightly different way, parallelizing the streams, then you can win in two and a half times in comparison with the usual typical machine. This is from practice.
Therefore, the performance there actually turns out to be higher than on the local machine due to the fact that you implement this multithreading or, in other words, the MapReduce approach, when we parallelize the stream and then finally only collect. That is, only the flow of the assembly occurs synchronously, and the load distribution should occur horizontally. But this is a zone of architecture. Approximately the same and 1C cluster works in 8.3. All code must be asynchronous, must be executed in several threads so that it can be parallelized, memory managed, and so on.
You say that 1C is naughty when they say they know only 1C, because besides this they also know other programming languages. Does this mean that we are on the verge of a turning point? Previously, 1C-nickname was the one who programmed in 1C. Now, he should understand virtual machines, know XML, understand how to configure Postgre servers for Linux ...
What you are talking about is called a "Performance Expert". He likes to understand how 1C behaves in a Linux environment or in a Windows environment. He studies performance on different servers - on Windows Server 2012, Windows Server 2008 - they cost different money. Some of them are more convenient for him, he can work with it. He does not do business automation at all, but still he is a 1C specialist. This facet is leveled. He is just a specialist, an expert in infrastructure, environment for 1C.
You put such a person, for example, a Tomcat application server in Java, and he says: “Oh, some kind of regular servant, not 1C-ny, well, yes, in Java, but behaves the same way”. And he suddenly becomes an application server infrastructure support specialist.
Do I understand correctly that, unlike the IT specialist of the 95th year, who set 1C himself, and then in two days gave the business some half-working, but working solution, the current 1C-nick, in the narrow sense, plays the dominant role, and the remaining 2 development department perform a supporting function? Does it not happen that due to this we come to “world peace”?
Maybe so, but there is another question. Again in the money. God bless him, with technocracy, technical issues are all settled there. Now the question is how to monetize it and how to make money on it.
So now, when switching to the “service model”, a new fashionable topic begins - these are microservices, microcommands. People start to talk about the fact that IT suddenly becomes not a cost center (as a support service), but a center of earnings. According to the new ideology, any IT-service should earn, therefore, in fact, we should hang up the billing (benefit calculation) for internal service, and conditionally 1C-nick buys some of its services from the dotnetchik.
In fact, some internal cost accounting.
Yes, and if each microprocess there will make a microinstruction, then they simply buy these services from each other on partnerships. Moreover, the model of franchising and model “update” is also just one of the services. For a business customer, he looks like a typical 1C-nickname, who arrives, updates ITS, can talk heart to heart with an accountant, tell him about the innovations that he read on ITS. This is a service, support for basic money. But this service also provides project activities, because if you have a service for such support, then the development department also develops as a whole as well, because the actual requirements of the business customer are actually brought by the employees of this service. with users heart to heart talk.
Can we say that this is, in fact, some kind of technical support?
It is impossible. Technical support are those that techie and do nothing else. In reality, this is not the case.
The following happens: the person went, updated the ITS, spoke with the payroll accountant, and returned back to the development team. At this point, the accountant sent for the development of the requirement to add props "Daisy" (this is a real requirement, it really sounded like that). This developer looks and indignant: what a props "Daisy", what is it all about?
And here that “renovator” comes to them, and explains that, in fact, the problem is this: a large staff rotation, a person can quit and return next year, they are called “daisies” in the jargon. Because they have two employment contracts, and a certificate 2-NDFL one. Therefore, they asked to add such a requisite to the staff directory so that there were no problems with tax. It turned out that they talked about this "over tea" during the update of ITS. Consequently, it is not necessary to refine anything, because it is already in the standard ZUP functionality.
As a result, it turned out that this person made an analyst here, who traveled and talked.
But this again, provided that he knows the typical configuration well, and generally understands the essence of the problem, as well as the specifics of the business - that is, it must be 1C-nickname of a certain level.
Yes. But he must reach this level.
By itself.
The problem of people who do not develop can be solved in two ways. What is happening in the market now?
There are people like us and you who are developing themselves, realizing that there is no other way. On this occasion, there is Moore's law, which says that the further development of technology is faster and faster. So, we need to accumulate basic tools in order to keep up with development, to be faster than development. And then we can be in the trend and level these risks, because now we are thinking about what will happen tomorrow.
And there are other people who are already forced by the platform. You do know that many people left the 1C world just because they heard that 1C platform is bad and so on. But in fact, they could not restructure from approach 7.7 to approach 8, from approach 8 to approach 8.3. I often say that the approach of 8.3 in the Java world (and not only) is the de facto standard: a three-way call from the client and executed on the server. If someone from Java developers would come to 1C development now, then in 1C 8.3 he was on a horse, because here everything would be familiar to him
I remember when I switched from Delphi to programming in the seven, I probably programmed for several months as follows: first I formed the code in the English version in my head, and then translated it all into Russian code on 1C. And the error caught exactly the same way.
Well, these are the costs of the transition period. The goal is good.
Yes, the output was good.
And the second type of people are those who do not keep up with development (development goes on continuously). And as a result, they are either thrown out of the market, or substituted, or reclassified, or dismissed.
Either you are developing, or you are developing, or do you even fall out of this scheme?
Yes. You can generally leave, run away. In my memory there was a case, as an IT specialist went to the “wilderness, to Saratov” and became a farmer. He got tired of it, he realized that IT is not his. And this is also normal.
We say that any problem can be solved - sometimes the problem can be solved simply by “fixing losses”. If we understand that IT is not ours, we can sometimes (very rarely) fix losses and say - that's all, I am not an IT person.
This is in any field, in principle.
Yes, the word IT is removed, and in fact, there is a fixation of losses in any methodology. It happens, it's not scary.
I used to be an orthodox too. For example, when I was yaver, I was negative about 1C. When I was 1C-nick, I thought Java was too slow. But now, after I visited “on both sides of the barricades”, I came to a certain heterogeneous approach (to a certain “Zen”), that there is no “bad” and “good”, but there is a problem, there is language and there is a solution. The main thing that a person wants, what he feels love, desire, and so on.
We return to the fact that the priority is given to those people who love what they do. Their main motive is not money, but love for work and self-development.
They develop. They are usually at the hearing and see the result of their work. And there are those people who do not want to develop and perform some routine work. I will not say that I agree with their views, but such people are also necessary.
We are talking about development and about different types of people. As far as I know, you are a member of various communities, including 1C-s. We just talked about changing paradigms, concepts. In your opinion, will there be any change in the communities themselves? Will they somehow be rebuilt, for example, the same Infostart, will there be a change of the composition, the audience?
Infostart in any case will develop. Where it goes, it has been developing since 2006. I registered in the Infostart community in 2007.
We are the same age with you.
Before that, I spoke on 1C ProClub and 1C ++. I didn’t like Mysta very much, because there was little practice there, I didn’t develop there. It was probably possible to talk there, but this conversation did not develop me at all. And on the 1C ++ forum there was more professional communication and the authors of this component taught me a lot. Therefore, I communicated with them on the forum, learned from them, and they taught me.
By the way, do you know how we came to Infostart? Just at some point in 2007, we decided to publish our work there - then there was nowhere else to publish, and we liked the +1 rating system here. It was cool. We, as end users of the “plus signs”, were “measured” by them at that time :)
I do remember.
I have since observed that the InfoStart community is evolving, changing. It becomes less trolling, more professionalism, more opportunity to get an answer. Communication "about nothing" goes to social networks, and professionalism remains in professional communities. , , .
GitHub — . , « — ». , . , , . PHP – , «, , ». .. «», «».
.
Yes. , — , , . , .
, . – , , . , , . , , .
. , « ». SilverBulleters. – Silver Bullet. , , – . : ? - ?
. , .
, 4 : , , . 2 . . — . , . — « », , « ».
, , – « » . , , . , «- » . « — ». , , .
— ?
, . , , , . , – . – . . , – . . , .
, , . ?
- , , , , . . , , , .. , 3 , , .
— . . - , . . , , .
, . , . « » , . , — .
______________________________________________________________________________________________________
:
youtube
itunes
podfm
Presenter - Sergey Gorshenin
Dear friends, I welcome you to the “Ready Solutions” program. I represent our guest - this is a person with twenty years of experience in the field of information technology, IT specialist, technical director and just a good person. Our today's guest is Alexey Lustin.
Hello, Sergey. Hello colleagues.
______________________________________________________________________________________________________________________
Chronological reference
')
Aleksey Aleksandrovich Lustin was born in Anadyr in 1980 and studied at the Voronezh Technical University with a degree in CAD Engineering. Since 1995 - IT-specialist. He started out as a 1C programmer in the city of Lipetsk, was engaged in the automation of business processes in the Yuterra network. Since 2008 - leading specialist Java, Rails. In 2008 he moved to Moscow. From 2012 to 2014 - Deputy Head of the Department for Innovative Development and Systems Analysis at the Svyaznoy Group of Companies. Currently he is a technical director, trainer, architect-consultant of the Silver Bullet company.
Alexey, for a start, I suggest you to tell a little about yourself in order to designate what we will talk about in the future.
It is quite difficult, like any beginning. In short, I speak about myself in such a way that I was born in Chukotka, far, far away, near the Arctic Circle. IT has been doing it since 12 years, that is, in fact, not 20, but in my opinion, already 22 years old. He has gone from the usual refilling of cartridges, the operator for filling out incoming invoices, to the architect, technical director and so on.
How did I go through it? Here, probably, it is necessary to tell longer.
I came from Chukotka in 1997 to go to college, because the main group of good teachers was concentrated in the central region, and not in the Far East. Maybe even in Novosibirsk, but in general, at that time, high-quality IT education could be obtained in technical cities - either in Moscow, or in Voronezh, or, of course, in St. Petersburg. Unfortunately, this was not the case in the Far East. To provide myself with money, and the time was hard, I took on any work on IT. This is the administrator of a computer club, and a collector of computer equipment, and the developer of 1C, oddly enough, even in the old-old version. I also tried my hand as a Java developer, when it was just beginning to develop, and as a Clipper database developer, who still remembers such things. For those times it was fun enough.
And after the end of the student phase, serious life had already begun, and there I went a different way: I was both a programmer and a project manager, depending on what the global goal of the employers I worked for (business owners, IT directors or IT executives in general).
Well, and then, in Lipetsk, I worked for the company Yuterra - this is a fairly large Russian holding company that combines a network of non-food goods stores - at that time there were about 112 supermarkets from Blagoveshchensk to Kaliningrad.
A little later, in order of my own development, I moved to Moscow on my own initiative, closer to the grandees of our industry, in order to really gain experience and share my approaches. So I got to Svyaznoy and worked there for about 4 years. I was there both a leading programmer and an internal technical leader; last year my role in the company was called “1C Competence Center”. So you understand, there are about 200 specialists in 1C.
Well, knowing the Messenger, I understand the level, yes.
Yes. Here is a way of life. In terms of my personal relationship with him.
IT paradigms from the four authors
So, in the 95th year you started as an IT specialist. Do you think that the model of an IT specialist of the 95th year differs from the IT specialist of 2015?
In general, nothing has changed, in my opinion.
But what about the changes in paradigms, in concepts, in technology development?
Paradigms have not changed. All the paradigms on which we are now parasitic, invented by the four authors - this is Brooks, Wirth, Ward and Hyudak. In fact, they invented these paradigms in 93-94 years (I am not talking about software engineering, which was invented in the 70s). Everything that is happening with technology is laid there. Nothing changes. On the other hand, only the practice of applying these paradigms is growing, that is, an IT specialist who started in 95th year has, conditionally, 20 years of experience in applying these paradigms. This is his only competitive advantage over a person starting in 2005. Just a person starting in 2005, has 10 years of practice. That's the whole difference. And the paradigms are the same.
Good. And let's remember the 95th year. Still, IT people then - these are people who loved IT, they came to this field at 12, at 15 years ... It is no secret that now some of IT people choose IT only because it is fashionable.
Yes, now really big fashion for IT. But then they chose, because it was fashionable. For example, you can recall the “dotcom boom” of 2000, which affected Russia, because even then there was a fashion for IT. Yes, and 3 years before, in '97, the fashion for engineering specialties was also very large.
I will explain why. In Russia at that time there was a general lack of IT specialists of any nature. For a good IT-specialist were willing to pay big bucks. They were no longer willing to pay for lawyers, economists, but they were ready for IT. It was quite an interesting time. If we recall that in the 93rd year, the passing score for engineering specialties was 9 - three triples, then in the 97th year for engineering specialties related to IT, the passing score was already around 14-15 - one four, two fives. This is an indicator. And, in my opinion, this passing score is still not reduced. That at the beginning of the two thousandth, that now the passing score and requirements for IT-specialties remain at the level of 97, 98, 99-ies.
On the other hand, and then in IT they went, including for fashion. Therefore, now many of those who started back then are no longer IT specialists. Some of them are businessmen, someone became an economist, someone left.
There were not those who came for fashion, but those who really loved it. And if we see an IT specialist with ten years of experience, then he most likely went through all the crises of this growth - 1, 3, 5 years. As in family life, you married IT, then a year later you realized who you married, after 3 years you had a falling out, and after 5 years you finally reached a certain “Zen” when you understand that you are from each other need to.
Such a philosophy.
You have a rather interesting experience, but you cannot say that you are an IT specialist, although you yourself insist on it. Why?
I have two theses that reflect my experience and the general concept of what I am for a specialist: this is Overskilled — the so-called “skill overkill” (there are too many), and at the same time Troubleshooting is a “person - problem solver”. I usually say that I can solve any problem in IT. Due to the experience, due to the technique, due to the developed algorithms - in general any. Because any problem can be solved. It doesn't matter if this problem is related to 1C or not. There can be no unsolved problems. This is the first.
Since I live with such an approach for a very long time, can you imagine how many problems I have already solved? That is why, in fact, I am overkill on skills. I can solve the problem with Ruby, with C ++, with the Assembler I can even, if there are problems. With 1C, of course, with business applications, with applications written by third-party developers, with OpenSource, Commerce applications. I have just experienced it all. Cartridge accounting? Yes please.
“CANNOT BE UNSOLVED PROBLEMS”
Business Application Specialist
Then I have a counter question. My career began in the distant two thousandth, and I started exactly with 1C. But after 6-7 years, I stopped being a single member in the truest sense of the word, although, probably, somewhere deep down in my soul, I still program at the seven. But why are you with such a rich technical experience still consider yourself to be one?
In fact, I usually call myself a business application specialist. And in Russia, and not only in Russia, it is impossible to create business applications without the 1C platform.
Can we say that in this case 1C is the most common platform for creating business applications, at least in Russia?
1C is an implementation of the concepts of a problem-oriented language. Let's go back to the 95th. There was a problem. There are a lot of businesses. Imagine how many configurations are now in Russia, written from scratch, without taking into account the typical. That is, de facto, we can say that every 1C specialist wrote his configuration from scratch. Bad, good is not important. On the INFOSTART.RU portal, 450,000 active users, 1C specialists are now registered, that is, we can say that each of them has written his own configuration.
“1C IS THE IMPLEMENTATION OF THE PROBLEM-ORIENTED LANGUAGE CONCEPTS”
You know, I will not go far, for example, I have 4 full-fledged software complexes. This is not counting the minor developments and reports.
I am usually a supporter of saying that how many 1C specialists are, so many atypical 1C configurations written from scratch. And how many businesses in Russia? They are much more.
Probably all the same business models?
Yes, business models. Most often, the business architecture is directly proportional to the architecture of the team creating this business. There is no the same business and the same process also does not happen. Even if they are based on the same principles, the concrete implementation is always different. We must accept this.
Where did the 1C platform and its language for developing business applications come from? All this appeared in response to the problem that the business grew in Russia - there was a transition from the old system to the new: many small enterprises appeared, the growth was very rapid. There were many small businesses. They could not all work on "template" business practices ("bestpratiksam").
Then there were no such practices as such in Russia.
Yes, there was no consulting as such for this market yet. But it was necessary to somehow work.
There is a need for a language to quickly describe business processes. Because you will not find in any platform a language in which you can write “Documents. Entry Bill. Pass ()” (register, print, send to printer, read off sale and so on). And it had to be described.
Conventionally, the idea was that even an accountant would be able to describe his business process. It didn’t quite work out - anyway, special people were needed who had an algorithmic base and could describe business processes in a special DSL language (problem-oriented language). But this way it was possible to reduce costs and quickly automate any business from 5-10 people.
In essence, this saved time and expenses for the creation of these base classes, base models, because they were already in fact.
Yes, it was a huge leap. The fact is that at that time, in the 95th year, the main language for developing business applications was C ++, which had a low level of abstraction.
Then on Clipper a lot was written, as I recall.
Clipper was too low-level language. Then Java just appeared, as a kind of framework that raised the level of abstraction. But he raised this level only up one stage - he allowed the component model to build, and 1C jumped right up through 5 levels, immediately to the level of a businessman. Even now, after 20 years, neither abroad nor in Russia there is still no such language that would allow to describe business processes directly.
Now banks are trying to create a language for automating traders and describing their algorithms in a certain DSL language - they are trying to do this in the Java world (or, more precisely, in the functional world). And we have such a language, conditionally, has existed for 20 years, so we are in a favorable position here. And it is unlikely that a language of the same level will appear there in the near future, making it possible to level up the system’s work at basic levels so effectively that at the first stage it is possible not to dwell on problems with the code, memory, so that you can immediately think about the benefits of business.
Remember, why did you come up with Agile? In order for developers on third-party platforms to start thinking about the benefits for the end user. They didn’t do space memory management algorithms, but thought how these algorithms can be useful for business. Not a businessman in a shirt (the owner, who invests money in it), but the end user - the storekeeper, the operator, the accountant.
There is still such a moment: how to justify for a business the need to optimize some algorithm in order to win one and a half megabytes of memory? In C ++, it is almost impossible to prove the feasibility study of such optimization for a business. And with this it’s easier, his tasks are close to business goals, so he can explain what will become faster, what will improve (and he can implement the optimization from memory just with C ++).
And if you go back to why I consider myself a 1C specialist - I always work with businessmen. Initially, 20 years ago, I started at the Mandatory Medical Insurance Fund. My first task was to automate the collection of payments on CHI. My first customer was the head of the OMS Fund, he explained the strategic business goal: why does he need it and so on. And the technical implementation and the technical risks associated with it were completely on my side - the choice of platform, whether it would be Access, Clipper, and so on, completely fell on me.
This was my first task - at that time I was 15 years old and there were no specialists in our city in databases at all. Besides me and Andrew, my colleague is a good C ++ programmer. And in our team there was a “senior comrade”, a radio technician of 35 years old - he explained to us how to implement everything technically, and then “combed” what was done. And we, the young guys, took the strategic goal of the person who was responsible for all district compulsory medical insurance, and tried to implement it. Bad, good, with errors - it doesn't matter anymore. Probably, if he sees us now, he will say that we all did badly there. But it normal.
Is it possible to say that it was for you that you decided then that you need to develop, that by setting your goal at the head of your work, he gave you an impetus to further development?
Yes, of course, then we were strongly influenced by what he believed in us. We were young tenth-graders, we won at the Olympiads, such IT specialists to the bone marrow - but, in fact, we were boys with no work experience. And he took a chance and gave us the opportunity to try to do something. And we tried, yes. In order to be close to computers and generally be close to everything. It was such a good motivation.
Different levels of abstraction
I remember the beginning of my career - it was Vodokanal, there was 1C. I found 1C under DOS version 2.0, then moving to 7.5, to 7.7. But when I started writing something under 7.7, I very quickly realized that I was missing the very accounting background. I then went to a three-month course, and this, for me, probably, on the one hand, served as an impetus for further development, and on the other hand, it served as a kind of click to get me to start moving from programming further into the depth of business processes. Is it possible that IT-specialists who write programs or applications for various business tasks will move to get away from programming and move to adjacent planes, to other platforms, for other tasks or to another area in general?
Yes, it usually happens. The world is divided. At first you just have some programming skills. You can write on Pascal, on C ++ something, some algorithms. This is a basic skill - at this stage you have no experience in working with business problems. But you get 1C-platform. You try to solve business problems with it, pass it through yourself and your first level appears. And what next - you decide. You can go to businessmen - it happens. You might just realize that IT is not yours.
You can go to business methodologists (or else this is called an analyst-methodologist). This is not the one who describes the methodology, but the one who is trying to be a link between the developers (those who create a product) and business customers. This is some kind of analyst who studies business processes, tries to find hidden requirements, and possibly attracts external methodologists.
Is this probably something close to the architect?
No, it is more related to system analysis. The architect - he builds component models, and this specialist is a kind of designer, he is more involved in analyzing exactly how to build the final model.
One way or another, you begin to specialize. You may want to leave the 1C-world at this moment, because suddenly you decide that you can design everything yourself on some low-level software. Take, for example, Java, as it is free, and you start to design something on it. As a result, people either begin to do something else, or, what happens most often, return back to 1C, but with different knowledge.
The amount of experience is directly proportional to the number of mistakes made as they grow. This is such a feature. Therefore, in the end, you still in most cases return to the 1C world. Maybe in a different capacity - for example, as a business customer, or as a chief architect, looking at 1C from the outside, but still working with her. Because 1C is a platform for automating business processes, and you don’t go anywhere on your own.
"THE NUMBER OF EXPERIENCE IS DIRECTLY PROPORTIONAL TO THE QUANTITY OF ERRORS"
But what about the reputation of one person? I remember when I started, many of my classmates laughed at me. Say, here we are programming in normal full-fledged languages, and you program in a language that is not enough that is made in C, it is also written in Russian.
Unfortunately or fortunately, I had a lot of teams in my work that are not the same. In general, I sometimes consider myself to be a specialist in Ruby on Rails and Java in addition to the 1C specialist - it so happens that I have experience writing applications on these frameworks. But there is another problem. People who write on .NET, Java, on anything, they are too far from the people. They do not automate the business. They write certain components for business, or rather components for those programmers who will then build a business application from these components.
In fact, they build cubes, from which further someone will build the final product.
Yes, this is how it works. After all, we never think about the fact that there are people who write programs in Assembler that implement the logic of microchips. And there, too, there are internal algorithms - various southern, northern bridges.
Come on easier. Here we control the mouse every day, we do not think about how the mouse movement is transmitted to the screen.
Yes, somewhere there are people who program microcrystals inside microcontrollers. There is a specialty "Physical Chemistry", where they learn to work at the level of microcontrollers, to program there. For them, 2 bytes is a lot. They have a memory there is not gigabytes measured, and bytes. These people give their finished components to other people who already describe working with devices at a higher level, for example, in C ++. Then some other people take these components, written in C ++, and make of them their applications, components, or libraries — already on .NET or on Java, or on something else. And then they come together, take these components and make them the ultimate business application. This is how it works.
Let's go back to the singles. Is it possible to say, based on your words, that a certain contradiction arises, that, on the one hand, you don’t want to be single member, but you have to.
You have to be one person, but you don’t have to be one person. This is the contradiction. For convenience of understanding, you just need to remove the word 1C. 1C is CJSC 1C, it is a company that develops, which invests both money and methodology inside this platform. You have to be a specialist in business automation, but you do not need to be a specialist in business automation. Resembles schizophrenia. But in general, you just have to learn how to switch the toggle switch.
That's when you communicate with a business customer, your task is not to discuss with him any technical issues - technical debts, technical architecture. The customer does not know how to build a cluster horizontally or vertically, how to build racks and so on. He tells you a certain wish. In general, the customer wants to get some of the alleged visual behavior of the mold as a whole. Unless, of course, this is not the customer who orders your compiler.
... Let's just say that if your customer is a business representative, not a representative of the IT industry.
Yes. We are talking about the final business. , . – , - – , , . .
IT-, , , -. . , 1 – , , ( , ) . , , , , .
. 1-. , , ?
. 1- , . , . , - , . 1- .
«1 »
, , ?
, , . 1 . , , . – , . , , . , , , .
80%, , 2 , 20% . .
. . , , 1: , , « », . 10 , , -, .
1 . . : - , 100 , , . , 1- . 1 . , , .
, ?
Yes. , . . 1- , IT, ( , ) — , -, . .
, , .
. 1- , , -. .NET- , IT, , .
L HTML (DSL, XML, WSDL, ..) language — .
«» , : ? , , , , . ?
-. -, , . , , .
10 . 1-, : , 1, 80% 1 - – Android, Clojure , . , . JavaScript, HTML. L HTML language — .
, 1. , , , XML 8 . -, WSDL (Web Services Description Language), L — . , 1, , , . .
, , , , , 5 .
5 , – - - , . , 6 , , , 20 , , . ( ) .
, , , . , , , Python, .NET, 1 XML, 1 HTML… , , 1 Web. , 1 - . - , 1, - «».
, , 2014- – , , . 1- (, , , , ). , « » — . . «», , .
, PHP , , . , , , , . Node.js Ruby on Rails, . , , - … .
«80% 1- »
, – «» 1-. , , , ?
– - . 1 «1», , . .NET, Java, Phyton . , 95- , , (, , ).
, . , – . , DSL , . . , : 1 .NET, — . .
. ( ) .NET Java 1- 5 . ( , , ) – 1-, , , .
, , , , . – .
, - ( ) : «, .NET, ». . , , : «, 1, ?» 1- .
: 1, , : « — .NET», .NET- : « 1, 1 — », . .
-
. , IT-. , , , , , 4 , 4 , , 2, , . , , IT-? , . , , .
, , , , .
IT- : , . IT- , – , IT- , , . . () — .
, ? , , — . , «», . ? , , .
. : , — , - . -, : « , » – , , .
IT- : . .
, , , . - .
« »
1-
. , 1-. ? 1- , ?
. : , , WMS – 7 . «», , Java, Ruby on Rails ++ . , : « (DZBZ2 Godex ), , ».
— 1, , – , ++, Linux. – . , … . , «».
. , . : , . , . , .
, , «», , WMS 1 . — . , , . – Java-, , , Java 1.
, 1 – , 1- - (, ..), .
, 1- Java-. – , .. , «1 – ». , , , -. , , WMS- 1 . , . . , – .
, .
, – , . , .
: , , , . : « ». : « -, , ». : « -? . 2 , . , . ?» , .
, , , -, ( ) .
I really liked how you touched the topic of costs. Recently, perhaps due to cloud computing, a situation has often emerged when the development team is being carried out somewhere in the regions. I, as an IT director, see in this, first of all, a reduction in costs. What are the pros and cons in this you see, precisely from the point of view of the technical director? Can you say that code quality suffers from this? Or is it all about the cost?
Quality suffers from this for sure.
Is the quality of the code unequivocally affected?
Everything suffers. First, when you have a non-distributed team, it’s almost impossible to manage. The only one who can manage this is, in my opinion, Alexander Belov. He has the only successful management experience.
The developers of this team should have a special psycho. You cannot hire a person with the nature of a stationary developer somewhere in Izhevsk or Yekaterinburg. You need to look for a completely different person who will be able to work in a remote team, reporting to the “center”.
In a distributed team, the costs of manageability are generally increasing. It seems that you save on money, on a payroll (payroll) due to the fact that you are going to the regions, but you start to lose controllability.
When you build your team, in any case you need to make a separate small development department methodological, detached, with your manager, with an office manager, and so on. Then it somehow works. Then this cost center pays off because it makes the final product. This is from practice.
If someone tells you that development in a distributed team will be more profitable, then they lie, they simply say that for money, according to payroll it will be less expensive. But the total cost — this is called total cost of ownership — is likely to be higher in remote development. And it's not about quality, but about the costs that you will bear.
That is, I correctly understand that they often say that direct costs - these initial investments (if you can say so) will be less, but about indirect costs that affect common bones - are they either forgotten or specifically kept silent about them?
Usually, when an office outside is taken out, people are aware of the situation with indirect costs, something like this: wow, it turns out, indirect costs are much higher than in Moscow, St. Petersburg or other central cities.
Is there a difference between simple developers and 1C developers from the point of view of remote companies?
From practice I have not seen. Money is different, but the interest is the same.
At the remote office, you still need a “champion”, the cost of which will be two times higher than the typical average salary level. And in the end, you will need to compensate for this small leadership at the expense of office costs, maybe even reduce the costs of marketing, accounting. But sometimes an accountant will need a local just in case, or an office manager. Practice such.
Well I myself come from the regions. I’ve been in Moscow only for the last 4 years, and before that I had shook from Rostov to Pskov and even in Siberia, so it happened that I was for a while. I mainly worked in the regions.
About clouds
Good. And if you use any modern tools for group development in the clouds, where everyone works in one place and does not communicate live?
First, for 1C of the world, this has not yet been created. There are our attempts in collaboration with the guys, specialists in Linux, now to deploy the cloud configurator - a special system for group development of configuration in the cloud with a customized environment, with all the tools. This configurator is currently undergoing internal beta testing. Maybe it will be possible to roll it out in the near future or the 1C company itself may possibly do something similar in the future.
Do I understand correctly that this is just a kind of virtual machine running the configurator, then the question is how is multiple access organized there, how is the joint development in it going?
So let's separate the flies from cutlets. We remove the word 1C. Infrastructure developer in the cloud - this is a big problem now in the entire market. In 1C, in Java, in .NET, in Java Script and so on.
Now there is such a concept Vagrant. 1C: Bitrix uses it (just started using it). This is when a little virtual tuned machine comes to you with a deployed Bitrix - so that you don’t waste time, but immediately start coding. In the same way, you can deploy Bitrix locally on your virtual machine, or somewhere in the cloud - yours, if you have one, or in a private one - whatever. As a result, we operate with these small virtual operating systems that come preinstalled to you and are automatically updated. Such a trend is everywhere. Therefore, the word 1C is removed - this is everywhere now.
This reduces costs. For example, to buy iron. Because if we imagine some ideal stationary computer of the 1C developer, then it will now cost, probably, about 40 thousand rubles, if you buy it very, very cheap. Although at current prices even, probably, already 60. There should be something like i7 (multi-core), 16 gigs of RAM and a terabyte SSD screw to work with the configurator cache so that everything fits.
Well, yes, software requirements are growing faster.
Yes, and it is necessary to recoup 90 thousand rubles. If you work in Moscow, you can somehow do this, and in the region 90 thousand rubles, even with a loan, is quite difficult to recoup.
And when using a remote environment, you just need a browser or some stationary client that will work with it over the network. With the development of communication channels (and they are still developing), this will be quite good and convenient. Therefore, it is a really interesting trend.
And how in your case with the cloud configurator will the speed of the channel be affected? It seems clear that this will affect the display of the picture, but how good will the performance on the cloud side be? And what will it depend on?
From practice - the performance there will be 3-4 times higher than, say, on your local machine, even with an SSD. Because the performance of the solution itself is already dependent on the supplier of this solution.
All of our final business solutions, final consulting, practices, we still sell. But the tools for people to develop, we lay out in OpenSource.
How was our community formed? We shared our tools, our work. Yes, and Infostart was so formed. All tools are laid out. So we, it turns out, are debugging it. The toolkit (the same Snegopat), is in many ways more productive than 1C, several times. Working with the configuration, working directly with the 1C storage is also implemented there several times faster than regular work. Due to the fact that there are good C-shniki, assemblers sometimes come across.
Well, yes, Mercedes AMG, he, of course, better than just a Mercedes, because it is brought to mind.
Yes, for example, there is an opinion that 1C on virtual machines slows down and reduces the performance by a factor of two. But there is another opinion that if properly configured virtualization, it will not lose in performance. And if you correctly configure the parallelism of the threads, that is, to implement the code of the internal tooling in a slightly different way, parallelizing the streams, then you can win in two and a half times in comparison with the usual typical machine. This is from practice.
Therefore, the performance there actually turns out to be higher than on the local machine due to the fact that you implement this multithreading or, in other words, the MapReduce approach, when we parallelize the stream and then finally only collect. That is, only the flow of the assembly occurs synchronously, and the load distribution should occur horizontally. But this is a zone of architecture. Approximately the same and 1C cluster works in 8.3. All code must be asynchronous, must be executed in several threads so that it can be parallelized, memory managed, and so on.
About experts in the world of 1C programming
You say that 1C is naughty when they say they know only 1C, because besides this they also know other programming languages. Does this mean that we are on the verge of a turning point? Previously, 1C-nickname was the one who programmed in 1C. Now, he should understand virtual machines, know XML, understand how to configure Postgre servers for Linux ...
What you are talking about is called a "Performance Expert". He likes to understand how 1C behaves in a Linux environment or in a Windows environment. He studies performance on different servers - on Windows Server 2012, Windows Server 2008 - they cost different money. Some of them are more convenient for him, he can work with it. He does not do business automation at all, but still he is a 1C specialist. This facet is leveled. He is just a specialist, an expert in infrastructure, environment for 1C.
You put such a person, for example, a Tomcat application server in Java, and he says: “Oh, some kind of regular servant, not 1C-ny, well, yes, in Java, but behaves the same way”. And he suddenly becomes an application server infrastructure support specialist.
Do I understand correctly that, unlike the IT specialist of the 95th year, who set 1C himself, and then in two days gave the business some half-working, but working solution, the current 1C-nick, in the narrow sense, plays the dominant role, and the remaining 2 development department perform a supporting function? Does it not happen that due to this we come to “world peace”?
Maybe so, but there is another question. Again in the money. God bless him, with technocracy, technical issues are all settled there. Now the question is how to monetize it and how to make money on it.
So now, when switching to the “service model”, a new fashionable topic begins - these are microservices, microcommands. People start to talk about the fact that IT suddenly becomes not a cost center (as a support service), but a center of earnings. According to the new ideology, any IT-service should earn, therefore, in fact, we should hang up the billing (benefit calculation) for internal service, and conditionally 1C-nick buys some of its services from the dotnetchik.
"NEW FASHIONABLE THEME - MICROSERVIS, MICROCOMANDS"
In fact, some internal cost accounting.
Yes, and if each microprocess there will make a microinstruction, then they simply buy these services from each other on partnerships. Moreover, the model of franchising and model “update” is also just one of the services. For a business customer, he looks like a typical 1C-nickname, who arrives, updates ITS, can talk heart to heart with an accountant, tell him about the innovations that he read on ITS. This is a service, support for basic money. But this service also provides project activities, because if you have a service for such support, then the development department also develops as a whole as well, because the actual requirements of the business customer are actually brought by the employees of this service. with users heart to heart talk.
Can we say that this is, in fact, some kind of technical support?
It is impossible. Technical support are those that techie and do nothing else. In reality, this is not the case.
The following happens: the person went, updated the ITS, spoke with the payroll accountant, and returned back to the development team. At this point, the accountant sent for the development of the requirement to add props "Daisy" (this is a real requirement, it really sounded like that). This developer looks and indignant: what a props "Daisy", what is it all about?
And here that “renovator” comes to them, and explains that, in fact, the problem is this: a large staff rotation, a person can quit and return next year, they are called “daisies” in the jargon. Because they have two employment contracts, and a certificate 2-NDFL one. Therefore, they asked to add such a requisite to the staff directory so that there were no problems with tax. It turned out that they talked about this "over tea" during the update of ITS. Consequently, it is not necessary to refine anything, because it is already in the standard ZUP functionality.
As a result, it turned out that this person made an analyst here, who traveled and talked.
But this again, provided that he knows the typical configuration well, and generally understands the essence of the problem, as well as the specifics of the business - that is, it must be 1C-nickname of a certain level.
Yes. But he must reach this level.
By itself.
About the transition period
The problem of people who do not develop can be solved in two ways. What is happening in the market now?
There are people like us and you who are developing themselves, realizing that there is no other way. On this occasion, there is Moore's law, which says that the further development of technology is faster and faster. So, we need to accumulate basic tools in order to keep up with development, to be faster than development. And then we can be in the trend and level these risks, because now we are thinking about what will happen tomorrow.
And there are other people who are already forced by the platform. You do know that many people left the 1C world just because they heard that 1C platform is bad and so on. But in fact, they could not restructure from approach 7.7 to approach 8, from approach 8 to approach 8.3. I often say that the approach of 8.3 in the Java world (and not only) is the de facto standard: a three-way call from the client and executed on the server. If someone from Java developers would come to 1C development now, then in 1C 8.3 he was on a horse, because here everything would be familiar to him
I remember when I switched from Delphi to programming in the seven, I probably programmed for several months as follows: first I formed the code in the English version in my head, and then translated it all into Russian code on 1C. And the error caught exactly the same way.
Well, these are the costs of the transition period. The goal is good.
Yes, the output was good.
And the second type of people are those who do not keep up with development (development goes on continuously). And as a result, they are either thrown out of the market, or substituted, or reclassified, or dismissed.
Either you are developing, or you are developing, or do you even fall out of this scheme?
Yes. You can generally leave, run away. In my memory there was a case, as an IT specialist went to the “wilderness, to Saratov” and became a farmer. He got tired of it, he realized that IT is not his. And this is also normal.
We say that any problem can be solved - sometimes the problem can be solved simply by “fixing losses”. If we understand that IT is not ours, we can sometimes (very rarely) fix losses and say - that's all, I am not an IT person.
This is in any field, in principle.
Yes, the word IT is removed, and in fact, there is a fixation of losses in any methodology. It happens, it's not scary.
I used to be an orthodox too. For example, when I was yaver, I was negative about 1C. When I was 1C-nick, I thought Java was too slow. But now, after I visited “on both sides of the barricades”, I came to a certain heterogeneous approach (to a certain “Zen”), that there is no “bad” and “good”, but there is a problem, there is language and there is a solution. The main thing that a person wants, what he feels love, desire, and so on.
“THERE IS A PROBLEM, THERE IS A LANGUAGE, THERE IS A SOLUTION”
"For the love of art"
We return to the fact that the priority is given to those people who love what they do. Their main motive is not money, but love for work and self-development.
They develop. They are usually at the hearing and see the result of their work. And there are those people who do not want to develop and perform some routine work. I will not say that I agree with their views, but such people are also necessary.
We are talking about development and about different types of people. As far as I know, you are a member of various communities, including 1C-s. We just talked about changing paradigms, concepts. In your opinion, will there be any change in the communities themselves? Will they somehow be rebuilt, for example, the same Infostart, will there be a change of the composition, the audience?
Infostart in any case will develop. Where it goes, it has been developing since 2006. I registered in the Infostart community in 2007.
“INFOSTART WILL DEVELOP IN ANY CASE”
We are the same age with you.
Before that, I spoke on 1C ProClub and 1C ++. I didn’t like Mysta very much, because there was little practice there, I didn’t develop there. It was probably possible to talk there, but this conversation did not develop me at all. And on the 1C ++ forum there was more professional communication and the authors of this component taught me a lot. Therefore, I communicated with them on the forum, learned from them, and they taught me.
By the way, do you know how we came to Infostart? Just at some point in 2007, we decided to publish our work there - then there was nowhere else to publish, and we liked the +1 rating system here. It was cool. We, as end users of the “plus signs”, were “measured” by them at that time :)
I do remember.
About IT Communities
I have since observed that the InfoStart community is evolving, changing. It becomes less trolling, more professionalism, more opportunity to get an answer. Communication "about nothing" goes to social networks, and professionalism remains in professional communities. , , .
GitHub — . , « — ». , . , , . PHP – , «, , ». .. «», «».
.
Yes. , — , , . , .
, . – , , . , , . , , .
« , »
. , « ». SilverBulleters. – Silver Bullet. , , – . : ? - ?
. , .
, 4 : , , . 2 . . — . , . — « », , « ».
, , – « » . , , . , «- » . « — ». , , .
« , – « » »
— ?
, . , , , . , – . – . . , – . . , .
, , . ?
- , , , , . . , , , .. , 3 , , .
— . . - , . . , , .
, . , . « » , . , — .
______________________________________________________________________________________________________
:
youtube
itunes
podfm
Source: https://habr.com/ru/post/254493/
All Articles