Over 40 years of his career at CERN (International Nuclear Research Laboratory located in Geneva), Rene Brun has developed a number of software packages that have become widely used in High Energy Physics. For this fundamental contribution, he was recently awarded a special prize from the High Physics Particle Physics Division of the European Physics Society (EPS). We talked with him about the key events of this story.
It is hard to imagine that the same person can create so many of the most important and widely used software packages developed at CERN and in the field of High Energy Physics: HBOOK, PAW, ROOT and GEANT. This passionate and far-sighted person is Rene Bran, now an honorary member of CERN , who was recently awarded the special prize of the Department of High Energy Particle Physics EPS (Society of European Physics) " for his outstanding and original contribution to software tools for data processing, detector modeling and analysis, which helped create experiments in particle and high energy physics for decades . ” Over 40 years of his career at CERNhe has worked with various brilliant scientists, and we cannot forget that the realization of such efforts is always the result of joint efforts. Nevertheless, Rene had an undoubted merit in generating new ideas, proposing projects and working diligently and enthusiastically to turn them into reality.One of his creations, ROOT , is a data analysis tool widely used in experiments in high-energy and nuclear physics, at CERN and in other laboratories. ROOTalready went beyond physics and is now used in other scientific fields and even in finance. GEANT is an extremely successful software package developed by Rene Bran, which allows you to simulate physical experiments and particle interactions in detectors. Its latest version, GEANT4 , is currently the primary detector modeling tool.But, before the advent of ROOT and GEANT4 , which are well known even among the youngest physicists, they were offered many other projects and software tools. This is an exciting story that Rene told at a seminar organized at CERN by the Experimental Physical Department in 2017.It all started in 1973 when he was hired by the CERN data processing department to work with Carlo Rubbia (Nobel Laureate, one of the CERN directors) in the R602 experiment on ISR (the world's first hadron collider). His responsibility was to help develop a special hardware processor for the online reconstruction of collision models. But since this development was slow and did not take much time, Rene was asked to write software for reconstructing events in multi-wire proportional cameras. “At that time, I hated software, ” Rene admits, smiling. “I wrote software during my PhD thesis, studying in Clermont-Ferrand and working inCERN at the weekend, and I did not really like it. I joined the Carlo Rubbia group with a “promise” that I would work on the hardware, but very quickly I again became a “software guy” ... ”Soon Rene realized in software (programming on Fortran4) that they could not implement with using hardware, and in addition, he developed a histogram package called HBOOK. This made it possible to conduct a very simple data analysis, create histograms, fill them in and send the results to a linear printer. He also wrote a program called HPLOT, which specialized in drawing histograms generated by HBOOK.There were no graphic devices at that time, so the only way to visualize the histograms was to print them on a linear printer, and the programs were written in the form of punch cards.Rene fondly recalls the time spent on punching cards, not on the procedure itself, which was slow and rather tiring, but on the long gatherings that were in the room where the punches and printers of the Data Processing Department were located, as well as the cafe nearby ( at that time she worked around the clock with the sale of alcohol ). During these long hours, he could discuss ideas and new technologies with colleagues.Huge progress was made possible thanks to the introduction of teletype, which replaced card punchers. Users could create programs on disk and exchange data with a central machine called FOCUS, and at the same time, see on a roll of paper what they were doing, like on a regular type machine. “ The way it works can make people smile today ,” Rene recalls, “in order to enter FOCUS, you had to type a team that turned on the red light in the computer center. Seeing the light, the operator had to insert into the machine’s memory a tape of a connected person who could start a session on disk. When the user logged out, the session was reset to tape again. You can imagine the movement! But it was much faster than punch cards . "Some time later, the teletype was in turn replaced by the Tektronix 4010 terminal, which led to an even greater revolution, since it made it possible to display the results in graphical form. This new, very expensive device allowed Rene to accelerate the development of her software: first HBOOK, then another package called ZBOOK and the first version of GEANT. GEANT1 was created in 1974 by him, together with his colleagues from the Electronic Experiments group, and was a tool for modeling simple detectors using the Monte Carlo method. Gradually, they added features to this software and were able to play out particle collisions: that's how GEANT2 was born.In 1975, Renee joined the NA4 experiment, a deep inelastic muon scattering experiment that ledCarlo Rubbia . There, he participated in the development of new graphical tools that allowed printing histograms using a device called CalComp plotter. This machine, which worked with a 10-meter-long roll of paper, gave a much better resolution than linear printers, but was very expensive. In 1979, a microfilm system was introduced: histograms saved on film could be checked before being sent to the plotter to print only interesting ones. This reduced the cost of using CalComp .Rene was supposed to continue to work with Carlo Rubbia in the UA1 experimentfor which he did a lot of modeling. But instead, at the end of 1980, he joined the OPAL experiment , where he made all the modeling of the detector and created GEANT3 .While working on the HBOOK system, in 1974 Rene developed a memory management and I / O system called ZBOOK. This tool was an alternative to the HYDRA system, which was developed by Julius Zoll, a group of bubble chambers (also the author of another control system called Patchy).Considering that having two competing systems is pointless, in 1981 Emilio Pagiola suggested developing a new software package called GEM. While three people worked hard on the GEM project, Renee and Julius started testing together to compare their systems, ZBOOK and HYDRA, with GEM. As a result of these tests, they concluded that the new system was much slower than them.In 1983, Jan Butterworth, then director of the Computer Department, decided that only ZBOOK would be supported at CERN, and that GEM should be stopped, and the development of HYDRA was frozen. "My group leader, Hans Grote, came to my office, shook my hand and said: "Congratulations, Renee, you won." But I immediately thought that this decision was unfair, because in fact both systems had good functions, and Julius Zoll was an excellent software developer. “As a result of this decision, Renee and Julius started a collaboration and joined forces to develop a package that combines the best features of ZBOOK and HYDRA. The new project was called ZEBRA, due to the combination of the names of two original systems. "When Julius and I announced that we were cooperating, Jan Butterworth immediately called both of us to his office and told us that if after 6 months the ZEBRA system did not work, we would be fired from CERN. Indeed, in less than two months we were able to show a working primary version of the ZEBRA system. »At the same time, the histogram and visualization tools were under development. Rene has put together an interactive version of HBOOK and HPLOT called HTV, which runs on Tektronix machines. But, in 1982, the advent of personal workstations marked a revolution. The first personal workstation to appear in Europe, Apollo, demonstrated a leap in terms of performance and performance: it was faster, had more memory and a better user interface than any other previous device. “ Apollo invited me to go to Boston and visit them ,” says Rene. “ When I first saw the Apollo workstation , I was shocked. I immediately realized that this could speed up our development by 10 times. I forced myself to work and I think that in just three days I adapted about 20,000 lines of code for it. ”Rene's work on adapting HTV for Apollo workstation aroused interest among Rudy Beck, Luke Pape and Jean-Pierre Revol from the UA1 collaborationwho also suggested some improvements. Therefore, in 1984, three of them developed a proposal for a new package that will be based on HBOOK and ZEBRA, which they called PAW (Physics Analysis Workstation).
PAW Team: (left) Rene Bran, Pietro Zanarini, Olivier Cue (standing) and Carlo Vandoni .After the first period of uncertainty, the PAW project developed rapidly, and many new functions were introduced, including due to the increase in the memory of workstations. “At some point, PAW software grew so fast that we began to receive complaints from users who couldn't keep up with the development ,” Rene says, smiling. "Maybe we were a little naive, but definitely enthusiastic . ”The programming language commonly used for scientific computing was FORTRAN. In particular, at that time FORTRAN 77 (introduced in 1977) was widely distributed in the community of high-energy physicists, and the main reason for its success was the fact that it was well-structured and fairly easy to learn. In addition, very effective implementations of it were available on all machines used at that time. As a result, when the new FORTRAN 90 appeared, it seemed obvious that it would replace FORTRAN 77 and be as successful as the previous version. "I well remember the head of the computing division Paolo Zanella, who said: “I don’t know what the next programming language will do, but I know its name: FORTRAN . »In 1990 and 91, Rene, together with Mike Metcalf, who was a great FORTRAN expert, worked hard to adapt the ZEBRA package to FORTRAN 90. But, these efforts did not lead to a satisfactory result, and discussions arose about the possibility of continuing to work with FORTRAN or switching to another language. This was a period when object-oriented programming was taking its first steps, and also when Tim Berners Lee (creator of the Internet) joined the Rene group.Timhad to develop a documentation system called XFIND to replace the previous FIND, which could only work on IBM machines that were supposed to be used on other devices. However, he believed that the procedure he was supposed to carry out was a little clumsy and, of course, not a better approach to the problem. Thus, he proposed another solution with a more decentralized and adaptable approach, which required primarily a lot of standardization work. In this context, Tim developed his famous idea of servers and clients - the World Wide Web , developed using an object-oriented language ( Objective-C ). Once I asked Rene, “How was it?” His answer shocked me -“This was not thanks, but contrary to CERN.”It was a very hot period, because the design and simulation phase of experiments for the new LHC accelerator (Large Hadron Collider) began. It was important to decide on the programming language and software tools that will be used in these new projects.At the ERICE seminar organized by INFNin November 1990, and then at the conference on computer engineering in high-energy physics (CHEP) in Annecy (France) in September 1992, the “guru” of high-energy physics of the world gathered to discuss programming languages and possible directions for developing software in High Energy Physics for the Large Hadron Collider. Among the many proposed languages were Eiffel, Prolog, Modula2 and others.In 1994, two research and development projects were launched: RD44 to develop in C ++ a new version of GEANT (which will become GEANT4 ) and RD45 to study object-oriented database solutions for experiments at the Large Electron-Positron Collider LEP .According to Rene, his division was divided into three groups: those who wanted to stay with FORTRAN 90, those who relied on C ++, and those who were interested in using commercial products. “ I submitted a proposal to develop a package that would allow PAW to be transferred to the world of Object Oriented Programming. But the project, which I called ZOO, was rejected, and I was even offered to take a vacation, ” Rene admits.This blow, however, later turned out to be really good luck for Renee. Its leader David Williams invited him to join the NA49 experimentin the northern part of CERN, where someone was needed to help with software development. At first he refused. For many years, he led the GEANT and PAW projects and was engaged in modeling or software development for various groups and applications, so the consent to return to work in a particular experiment seemed to him a big limitation.But, he thought about it and realized that it was an opportunity to spend some time developing new software with complete freedom. He visited the NA49 building on the Privesan site and, seeing pine and squirrels from the windows, felt that this was indeed the very calm environment that he needed for his new project. Therefore, he moved his workstation from his office to the website of Privessan (“I did it on the weekend without even telling David Williams ") and, while working on NA49 , he learned C ++, translating most of the HBOOK software into C ++.In early 1995, Rennes was joined in NA49 by Fons Raidemakers, with whom he had already collaborated. They worked together a lot and created the first version of what became the famous ROOT system. The name comes simply from a combination of the initial letter of the email addresses of the two founders (René and Rdm, for Rademakers), the double O Object Oriented and the word Technology. But the meaning or word “root” was also well suited to be the basis for developing more software and using tree structures in its architecture.In November of that year, Rene held a seminar to introduce the ROOT system. “ The computer room was unexpectedly full! ” Rene recalls: “ I think this is because people thought that Fons and I had disappeared from the software arena and suddenly we were back again! ” And in fact, the ROOT system aroused considerable interest .But, while Rene and Fons were completely absorbed in working on their new software package, the RD45 project, which had the authority to decide which new software should be adopted in the new LHC experiments, suggested using the commercial product “ Objectivity". and much work continued to develop applications to meet the needs of High Energy Physics. According to Rene, there was a clear intention to prevent the development and spread of ROOT. In the spring of 1996, CERN's Director of Computing Lorenzo Foa announced that the ROOT project was seen as a private initiative NA49, which was not supported by CERN management, and that the official development line was aimed at Objectivity .“ I think that the LHC Computing Board did not have the right idea about the architecture of these software tools, so that it can be judged which solution was the best. Thus, they had to trust what they were told, ”Rene comments. "It’s always a problem when there is such a gap between experts - and users - working on something and people who have to make important decisions . ”However, Renee and Fons continued the development of ROOT and introduced new features, using the lessons learned from previous software packages (in particular, requests and criticism of users). In addition, they closely monitored the development of the official line with Objectivity to find out what the people using it were looking for and what the problems or difficulties were. “ The more we looked at Objectivity , the more we realized that it could not satisfy the needs of our community ,” adds Rene, “we knew that the system would collapse and that people would eventually realize it. This has given us even more energy and motivation to work hard and improve our product. »They received ongoing support from the NA49 and ALICE collaborations , as well as from many people from ATLAS and CMS who saw good potential in this software package. At that time, Rene collaborated with many people in both experiments, including Fabiola Gianotti (current director of CERN) and Daniel Frodevo, in particular, to model detectors. In addition, many users have trusted them for many years thanks to the support of PAW and GEANT users.The situation began to change when interest in ROOT grew outside of CERN (oddly enough, but the slogan Zagranitsa will help us! Works outside of Russia). In 1998, two experiments Laboratory them. Enrico Fermilab in America (Fermilab), CDF and D0 , decided to discuss the future of their software in light of the upcoming Tevatron session . Thus, they opened two tenders for software solutions, one for data storage and one for data analysis and visualization. Rene sent ROOT to both competitions. During the CHEP conference in Chicago (it was there that I first met with Rene personally, although I collaborated with him in absentia for quite some time.There he offered me a job) proposals were discussed, and on the last day it was publicly announced that CDF and D0 would accept ROOT. “ I did not expect this ,” Rene says, “ I remember that when the decision was announced, everyone turned and looked at me .” Soon, the RHIC experiments at Brookhaven National Laboratory made the same decision. When the physicists of the BaBar experiment in SLAC , after many years of trying to use Objectivity, realized that it was not as good a system as expected, so they switched to ROOT.Gradually, it became clear that the entire High Energy Physics community was “naturally” moving towards ROOT, the CERN leadership had to accept this situation and, in the end, support it. But this only happened in 2002. With the growth of the workforce allocated for the project (Rene then joked “ you are the first CERN employee who officially receives a salary for working on ROOT ”), ROOT continued to develop rapidly, and the number of users increased sharply. ROOT also began to spread to other branches of science and to the financial world. “ In 2010, we had an average of 12,000 downloads of the software package per month, and there were more visitors to the ROOT website than to CERN .”
Logo of the ROOT software package.
Rene retired in 2012, but his two most important offspring,ROOT and GEANT continue to evolve thanks to the work of many young scientists. “ I think it is important to have a constant incentive that will push you to improve your products and the emergence of new solutions. The contribution of young people is very important for this , ”comments Rene. But, as he admits, what really made him and his colleagues work so hard for many years is that there have always been several competitors, although in many cases they were a challenge and even interfered. “ When you have an adversary, but know that you are right, you are doomed to success . “Great attention to user requests was also very important because it helped develop software and build trusting relationships with people. "I have always said that you should give priority to user support, ”Rene explains. “ If you answer the request within 10 minutes, you will receive 10 points, in one hour you will receive 2 points, and in one day you will gain -10 points. Answers to questions and comments are fundamental, because if users are satisfied with the support you provide them, they are ready to trust what you offer next. "Now that he has resigned, Renee is still monitoring software development at CERN, but only as an external observer. This does not mean that he left aside his scientific interests, on the contrary, he now devotes most of his energy to a more theoretical project, since he is developing a physical model. In his free time, he loves gardening. He loves flowers, but cannot help but look at them with a scientific gaze: “ My colleague, a mathematician, and I have developed a mathematical model of how flowers are structured and grow .”Brilliant minds are always at work.
A little bit about yourselfROOT.
.
.
— . .
— « ».
, , « » ! .
, , , , .