Cuts and how to deal with it

TrV-Nauka publishes the thoughts of Yekaterinburg City Duma deputy Leonid Volkov ( lv@leonidvolkov.ru ) on how to identify the corruption component in the field of IT competitions and what needs to be done to reduce corruption in this area. The issues under consideration directly resonate with the problem of using 94-FZ in competitions for scientific research and development. Leonid Volkov is an independent IT consultant. Until recently, he led business development at a large software company, SKV Kontur.

Let's start with theory, and even, moreover, with definitions. Cutting is a competition, the main goal of which is not the creation (implementation, modification) of a software product, but the distribution of budget funds between the organizers and participants of the competition (in this case, the creation of a product may or may not be an insignificant side goal).

Corruption is the actions of officials aimed at providing illegal preferences to one of the participants in the competition in exchange for some kind of reward (kickback). Corruption is a much broader concept than a cut: for example, a kickback of 5% is not a cut, and simply creating conditions for “your” company to win is not a cut, but, of course, corruption.

Corruption factor - conditions conducive to the emergence of corruption relations. A corruption-inducing factor does not necessarily mean the presence of corruption, but the presence of a corruption-inducing factor should be punished in the same way. Just as a drunk driver on the way home from the bar will not necessarily hit someone, but this does not mean that you can drive drunk.

An indicator of corruption is events or facts that indirectly indicate that corruption may have occurred.

The topic of cuts in government procurement in the IT sector became very loud after Alexei Navalny achieved the cancellation of the Russian Ministry of Health’s competition for the development of a “social network” in 16 days for 55 million rubles. and the subsequent dismissal “at his own request” of the official responsible for this competition [1]. Soon after this, the head of the control department of the President of the Russian Federation, K. Chuichenko, reporting to D. Medvedev on the state of affairs with the state order, indicated that “at least 1 trillion” of the state order was being spent ineffectively (with a total volume of 5 trillion per year, Mr. Chuichenko postulated that such that the standard rollback percentage is at least 20%). In this regard, there was a list of industries “in which there is a lot of theft,” which included the IT industry. I think this is directly related to the wave in the blogosphere that was generated by discussions of the Ministry of Health and a number of subsequent competitions.

Let's figure out whether there are so many cuts and corruption in IT competitions, how to identify them and how to fight them. The problem is that, without experience in developing large, serious, loaded information systems, without experience in managing a large team, in which, in addition to programmers, there are many representatives of related specialties, without understanding the cost structure of the contractor company, it is impossible to assess whether it is or another IT competition cut or not. But even for a specialist, it is possible to estimate the project budget using a well-written competition documentation (CD) with an accuracy of perhaps plus or minus 50% in one direction or another. In other words, if the design documentation states a contract price of 10 million rubles, then an experienced specialist, having spent a couple of hours studying it, may come to the conclusion that “the price is clearly too high, since there is no more than 5 million of work involved” or “the price is clearly underestimated, since it’s not worth taking for less than 15 million.” If a person says that he can distinguish between a project worth 7 and 8 million based on the design documentation, then this person is a charlatan. Therefore, according to the documentation, it is possible to identify only very obvious and blatant cuts, where the price is really inflated by three or four times. I don’t know how to distinguish a crystal-clear competition from a competition where there is a 20% kickback, and I don’t think anyone has such measuring instruments. This is the difference between IT competitions and, for example, the supply of potatoes to preschool institutions. How, say, to distinguish the developer’s profit of 30% included in the estimate from a profit of 10% and a rollback of 20%? This leads to the idea that the price of work is a very weak and unreliable indicator of corruption, even if it appears to be greatly inflated.

It is much easier to make a preliminary assessment not by cost, but by timing. It would be more correct to say that the accuracy will be the same, not very high, but what will be written in the design documentation is much more likely to not fall within the “confidence range”. In other words, looking at the design documentation, you can quickly estimate that this is a project “for 6-12 months” or “a year and a half, no less.” If the design documentation states “10 or 15 days,” this is a clear sign of cutting. This is how all nine cases included in the famous “Navalny list” were identified [2]. In seven of these nine cases (the first two are the supply of equipment, and there it cannot be said so categorically), the stated deadlines are unrealistically underestimated by 10-20 times and cannot be met.

Deliberately underestimated deadlines are very common and are a reliable indicator of corruption. But what exactly do they show? One thing. That the contractor is already known and the work has already been completed, and the competition only de jure consolidates the relationship between the customer and the contractor that developed de facto earlier in the year. This is definitely not good and illegal. But this in itself does not prove corruption. Real-life scenarios that may be hidden behind understated deadlines can vary over a very wide spectrum. At one extreme, unrealistically short deadlines are stated in order to cut off competitors; No one is going to make the product itself. There is a cut. At the other pole: the department has an IT office, which from year to year maintains the information system of this department for very modest money, and this fee does not increase; the relationship is built on trust, and when the department experiences interruptions in budget funding, it asks the contractor on his word of honor to wait, work for free, and will pay at the end of the year, when the budget money is available. Look, these are completely different situations, but they will manifest themselves to the observer in the same way - a competition at the end of the year with ridiculous deadlines.

Thus, underestimated terms indicate the presence of a corruption factor, which consists in choosing a supplier bypassing competitive procedures with the subsequent legalization of this choice through a competition in which competitors can no longer participate due to underestimated terms. A corruption factor is not yet corruption. Let's give an example: the Ministry of Education and Science of the Russian Federation planned to allocate money this year to finance Olympiads and sports programming camps for schoolchildren and students. It is absolutely obvious (to all specialists in the field of sports programming) that no one could apply for this money except universities that host the quarterfinal competitions of the World Championship (in fact, this money was mainly intended for holding these competitions). But money cannot be allocated otherwise than through a competition. Therefore, from the very beginning of the year, universities began not only to organize competitions and training for students and schoolchildren, but also to the painstaking work of writing applications for a competition that had not yet been announced. Of course, it was known in advance who would win and what amounts they would receive: it was simply about financing the Ministry of Education and Science of the Russian Federation for its own subordinate universities. But work on the competition documentation dragged on, and, as a result, at the time of publication it began to look completely indecent [3]. Summing up the results of the competition - November 30, reporting - December 15. During this time, universities were supposed to spend several tens of millions of rubles and supposedly train several hundred students. In fact, all this was successfully done during 2010, but the competition was canceled after all the scandals and posts of Alexei Navalny (along with a number of other competitions of the Ministry of Education and Science, including two competitions from Navalny’s list). It is unclear how universities, which at least spent a lot of effort on preparing documentation and applications, will get out of it.

At the same time, it would be wrong to completely reduce the whole problem only to the uneven functioning of the budget system and thereby try to “decriminalize” this corruption-generating factor. Yes, the budget system, in which if you don’t feverishly spend all the money at the end of the year, then next year funding will be cut, works disgustingly. But still, in IT competitions with short deadlines, the point is often not only that the money came up at the end of the year. The customer commits a crime - the choice of a contractor without a competition is usually carried out for other reasons.

The main consideration is that a normal technical specification for an IT competition can only be written by its future winner. The problem is that in our industry (and this is its specificity) creating a high-quality problem statement is often a more labor-intensive process than solving the problem. The classic Brooks ratio [4]: ​​allocate 1/3 of the project time budget to design and only 1/6 to development itself! (Another 1/4 is for testing modules and 1/4 is for testing the product as a whole. Of course, these recommendations from 1975 are partly outdated, but only partly.) The government customer is the department that wants to obtain an information system for itself - in 99% of cases he cannot, is not capable of writing a normal technical specification, because this is the most difficult part of the task. The statement of the task, as a rule, is written by the future contractor, and then, of course, he is interested in ensuring that the competition itself does not get out of his hands, and tight deadlines are the best protection. How is a contractor selected? Not on the basis of a competitive procedure, but on some other basis, including corruption (although not necessarily).

So, 94-FZ should ideally work like this: “The customer determined the scope of work - put it up for auction - the strongest and cheapest won - did it - everyone is happy.” In reality, in IT competitions it turns out like this: “The customer roughly described what he needed, agreed with the person who will do it, agreed on the price, wrote the design documentation, ensured the contractor’s victory, in parallel all this time he was doing the work - everyone is happy.” . In those rare cases when the customer nevertheless formulates the requirements himself, happiness, as a rule, does not come at all...

Is this scheme corruptible? Yes. Because corruption here is uncontrollable and the temptation for its occurrence is too great. What does the contractor pay for choosing him? How and why does the customer cut off others? How does a contractor convince the customer that he is the best? What tricks are used to ensure that someone “leftist” doesn’t win the competition? An envelope with money is very suitable for solving each of these issues, and sooner or later it arises. It is this set of circumstances that leads me to the idea that government money is cursed, and a development company that participates in competitions and applies for them is selling its soul to the devil. Ultimately, this inevitably leads to corruption.

So, drawing conclusions about corruption based on the combination of price and timing alone, without examining additional circumstances and identifying additional indicators, is unfounded. Another important indicator that allows us to judge the fairness of the competition is the persistence of its organizer in selecting the “right” contractor.

Another indicator is “mixing the tracks” behind the competition. For example, the well-known and previously very popular trick with Latin letters in the specification (so that the search would not find it), refusal to publish significant parts of the technical specifications, etc. Another typical technique is the inclusion of cut-offs in the design documentation, imperceptible to the uninitiated, discriminating against all “undesirable” participants . For example, the requirement for exotic software and hardware platforms or rare licenses.

Finally, there is another very important aspect that is often forgotten - the aspect of expediency. For example, I carefully studied the competition documentation for the third example from the “Navalny list” [2] called “Electronic office of the mayor of Moscow”, and it seems to me to be well written. The task was set well and clearly, and the stated price was 23 million rubles. - quite adequate to the volume of work. The question is different: in general, do taxpayers need such a software package? It seems to me - no. Creating a deliberately useless product, even for adequate money, should also be considered a form of cutting.

To summarize all of the above:

1. It is almost impossible to distinguish with the naked eye the presence or absence of a cut in a particular IT competition.

2. The main corruption factor in IT competitions is the predetermination of the contractor, due to the specifics of the technological process in the industry.

3. An abnormally short tender execution time clearly indicates that the contractor is known in advance, and this is a corruption-prone factor.

4. It is necessary to pay attention to the accompanying signs of corruption in order to understand whether corruption has manifested itself in this particular competition.

Constructive action program to combat cutting and kickbacks during GT competitions.

1. It is necessary to create a specialized structure - a single government customer in the IT field. The government customer must have a supervisory or expert council of respected industry representatives. The point is to ensure control over the quality of bids put up for auction, as well as to eliminate the corruption factor by making the customer an organization that is not directly interested in using the budget of a particular department.

2. All orders for the development, implementation, and maintenance of information systems in the interests of government agencies are placed exclusively through this single government customer. The single customer checks the task statements received from departments for adequacy and usefulness (an expert council helps him in this), and also translates the competition documentation from the language of the customer’s needs into the language of technical specifications for developers. Meaning: in a single customer we concentrate competent IT specialists, analysts and directors (who are simply nowhere to be found in departments) and with their hands we create intelligible productions without the participation of contractors.

3. Any order for development must be preceded by an order for design. The meaning: what cannot be defeated must be led. Namely, in IT we still cannot avoid the situation when a contractor writes technical specifications for himself, because this is the only way a competent technical specification can be written. Even our ideal single customer still won’t write all the technical specifications, but he will be able to check them!

4. A design competition is appointed by a single customer only if, based on the results of the design competition, it was possible to obtain a truly high-quality technical specification; the quality of technical specifications is assessed by an expert council and verified by the original customer (department). Meaning: the project is doomed to obvious failure if the technical specification is illiterate, contains ambiguities and ambiguities, and allows for different interpretations of goals and objectives, criteria for achieving goals.

5. The system of criteria for evaluating competitive applications may legally contain preferences. Preferences are provided to: the development company - during a competition for maintenance and implementation; to the development company or implementing company - during a competition for revision; the company that won the design competition - during the development competition

6. It is possible to postpone budget funding to the next year. Incomplete utilization of allocated budgets is not a reason for imposing penalties or cutting budgets.

Leonid Volkov

1. http://navalnylivejournal.com/514023.html

2. http://navalny.livejournal.com/514097.htnl

3. http://zakupki.gov.ru/Tender/ViewPur-chase.aspx?Pur ch asel d=899822

4. Brooks F.P. The Mythical Man-Month: Essays on Software Engineering. 1st ed. Addison-Wesley, 1975, p. 200.