β Scribed by J. Nadrchal
No coin nor oath required. For personal study only.
The Computational Physics Group of the European pared the facilities of contemporary symbol manipu-Physical Society in cooperation with the Faculty of lation systems and predicted that in future there would Sciences of the Comenius University in Bratislava, the be more effective methods for obtaining numerical Union of Czechoslovak Mathematicians and Physicists results from symbol manipulation programs and much and the Slovak Academy of Sciences organized a school better systems for symbolic integration. on computing techniques in physics from 5th until There were three lectures in the summer school 15th June 1975. that were devoted to numerical problems. The first The school took place in the Smolenice Castle about was the lecture by M. Roos from the University of 60 km north-east of Bratislava. There were 109 partici-Helsinki about data averaging and problems with inpants including 54 from the organizing country and consistent data; the second and the third were by 55 students and teachers from other countries. The F. James from CERN about function minimization program was not specialized, including numerical and Monte Carlo methods. The lecture about minimimathematics, computer science and hardware but all zation included a survey of methods most of which with close reference to physics.
are implemented in James' system MINUIT in The subject of the first lecture by K. Peckover from FORTRAN. the Culham Laboratory was the programming system C. Verkerk from CERN gave a very interesting lec-OLYMPUS well known to readers of this journal (see ture about special purpose processors. He showed that Vol. 7., no. 5). His second lecture dealt with techniques specially built hardware can be more effective for some in classical computational physics. Firstly the use of purposes than general purpose computers, e.g. procescontrolled numerical experiments to simulate the be-sors for fast Fourier transforms or for pattern classifihaviour of physical systems obeying classical laws was cation solve their tasks much faster than a general purdiscussed. Secondly the lecture dealt with the simplifi-pose computer. cation of computer applications in physics using sym-J. Nadrchal took as the subject for his lecture bolic programming.
"structured programming" and applied it to give rules Next G.A. Ososkov from the Joint Institute for for good programming. He analyzed some frequent Nuclear Research in Dubna (U.S.S.R.) discussed nu-mistakes in programming and proposed a procedure merical simulations and programming problems in for solving problems with the aid of computers. J. measuring systems in nuclear physics.
Moudry applied the topic "structured programming" A very important section of the program of the to the programming of control computers. school was a series of lectures on symbol manipula-As well as these lectures there were seminars given tion: J.A. Campbell from King's College, Cambridge, by the lecturers and other participants. Summaries of described LISP 1.5 and SNOBOL; A.C. Hearn from all lectures, in some cases the full texts, were prothe University of Utah (U.S.A.), the author of the pro-vided. (A few of the collections of summaries are still gramming system REDUCE, explained the differences available from the author of this report). between symbolic and numerical computing especially
The excellent lectures, the friendly and comforta. with regard to the needs of storage and time. He corn-ble surroundings and the enormous amount of work
π SIMILAR VOLUMES