Interviewer questions are bold, Dr. Zakharov's comments are in plain text.
Part b of this interview can be found here
When was Convekta started and how many employees did you have at that
time? When were the first DOS and windows versions of CA released?
Chess Assistant development was started in 1988 with the inspiration of Anatoly Karpov. We started with four programmers, all graduates of the Computer Science Department of Moscow State University (MSU). Although we had other duties, the first versions of CA appeared quickly. And it was very good from our perspective. So we decided to make it commercial. We were young and full of ideas. And it seemed to us that in a month or two, we would have the program ready to sell.
Then reality set in; month after month we felt that we were not making enough progress. Now I know that tuning the program for the needs of users, making the program interface intuitively clear, catching bugs, checking different hardware configurations, preparing the databases, writing the documentation, and even writing a setup program (in those days) - all demand a lot of time. So the first sales started only in 1991. Initially, we did not have a distribution network, so sales were slow. But even that small amount of money allowed us to create a company. Life in Russia was cheap those days. So hiring an employee for $50 a month was not a problem. The main expenses were for hardware and software. Having some money, we began to grow quickly.
And it still seems to me that CA DOS 2.0 was close to perfect for that time. By the way, the name CA 2.0 doesn't mean that it was the second DOS version, we named them 1.0, 1,1, 1,2, 1.3, 1.4 and then 2.0. Although CA DOS didn't have some of the functions that ChessBase had, a lot of people loved it for the speed and real chess tree. At the time, I thought that we would take a significant part of ChessBase’s market soon. But then we made a few serious errors.
I think it was 1994 when we started development of the Windows version of CA. We decided to rewrite all the code from the beginning, hoping to address some functional issues in the DOS version. Moreover, we decided to migrate from Borland Pascal to the C++ language. In parallel, we decided to develop chess teaching products. And the big mistake was that we had some free funds, so we decided to diversify, and we sought business as a software company.
Now, programmers know well that writing a Windows program demands much more effort than writing a DOS program (my estimation is 5 times more). And the problem is not that Windows is a bad system. It is a good system but it is very complex one. Programs should have a new level of visual design, should support many more functions, and should work in a multitasking environment. It was possible to make a good commercial product with one programmer under DOS. Windows raised the complexity of commercial programs much higher. So it is not surprising that some well-know DOS
programs were never adapted to Windows.
And what do you think happened then? After we wrote a lot of code in C++, a new windows version of Pascal (Delphi) appeared. After a comparison of C++ tools and Delphi, and much soul searching, we decided go back to Pascal, i.e. Delphi. However, part of the code related to database and searches remains in the C++ language.
Other employees in our company that were busy with non-chess software development simply went to a new business that was much more profitable than chess.
Our excellent teaching program, CT-ART 1.0 for DOS, was developed at this time, but it was not in the same niche as CA. So we needed another channel to sell it as well as good advertising. User feedback was excellent but sales of CT-ART were not what we expected (Editor’s note: Sales of CT-ART are now very strong). It seemed that not many people knew about the program.
As a result of our diversification and the extra development time taken for our windows version, CA 3.0 for windows appeared only in 1997. And it was not as good as we planned. Some functions that were in the DOS version, like searching for double games, were absent. The program was not stable. At the same time, ChessBase released their second Windows version. And we did not feel like CA 3.0 held up well in comparison. We lost a lot of users with this version. For other companies, this would have meant bankruptcy, but we survived. Moreover, I think that the lessons learned were very useful. We lost some good people. But the people that we have now are strongly devoted to chess. We have a new young programming team and a growing chess team. Our sales department grows too. So today we have the best team we've ever had - professional, young and devoted to chess. And we became a real company, not a group of enthusiasts.
Your organization has grown since then, how many people do you have now?
Now the Chess Assistant team includes seven programmers (including me), five chess masters working full time and a few part-time chess experts, there is a sales/shipment department that supports the web site also.
The programming team is a good mix of experience and youth. Two of them are still working on their studies, but it doesn't mean that they are weak in any sense. I hope that team will be larger very soon.
Part b of this interview can be found here.