ORGANISATION EUROPÉENNE POUR LA RECHERCHE NUCLÉAIRE CERN EUROPEAN ORGANIZATION FOR NUCLEAR RESEARCH SCIENTIFIC POLICY COMMITTEE. Eighteenth Meeting

Transcription

1 CERN LIBRARIES, GENEVA CERN/SPC/ October, 1960 CM-P ORGANISATION EUROPÉENNE POUR LA RECHERCHE NUCLÉAIRE CERN EUROPEAN ORGANIZATION FOR NUCLEAR RESEARCH SCIENTIFIC POLICY COMMITTEE Eighteenth Meeting Geneva - 4 November, 1960 TECHNIQUES OF DATA HANDLING FOR TRACK CHAMBERS

2 TECHNIQUES OF DATA HANDLING FOR TRACK CHAMBERS 1. THE FRANCKENSTEIN-IEP TECHNIQUE Active work on digitizing measuring machines for track photographs started at CERN in Early in 1957, contact was taken with a somewhat more advanced development in Berkeley. In their present shape, CERN's "Iep" instruments and Berkeley's "Franckensteins" are rather similar in scope and performance, each design possessing, however, some distinctive features of its own. During 1960 some continuous operating experience was gained with 3 Ieps which had teen manufactured at CERN. Three more (CERN-made) and a seventh - from the first batch made by an outside firm - will become available in With these machines, and working a normal day, it appears possible to measure in 1961 some 30,000 events. Track-measuring equipment, as it stands now at CERN, Berkeley and elsewhere, has by no means reached its final shape; improvements and adjustments are continuously suggested by practice and a standing technical force, distinct from the maintenance crew proper, has to be provided for that purpose. The total capacity of a "measuring factory" on the Franckenstein-Iep lines could, of course, be increased over and above the figure of 30,000 events per year. This can be done by increasing the number of machines and/or shifts, and by further spreading the geographical distribution of pictures to be processed (several universities in the member countries are acquiring measuring equipment which will increase the processing capacity available in Europe). Anticipating the effect of minor technical improvements, such an increase could go until some 10 5 events per year. Such a number, or further expansions for even higher numbers, would require a very large "industrial" undertaking, at the limit of or beyond what one would like to see in a laboratory in or around CERN. The problems of organization and communication would become far too unwieldy, not to speak of the material cost. For this reason, the tendency in many laboratories to-day is to look for systems basically more efficient than Franckenstein-Iep (in orier not only to transcend the expansion limit of Franckenstein-Iep, but also to bring down the manpower cost per event). Ir. CERN the need for such systems may become acute even before the 2 m chamber is put into operation. The need for new and more economic data handling would be imperative if and when experiments were performed in which simple events have to be counted with high-number statistics (several times 10 5 or 10 6 ). At present there exists no satisfactory measuring technique at all for such requirements. As long as measuring equipment is developed only on Franckenstein-Iep lines, such experiments will remain impossible.

3 Page 2 2. COMPUTERS a) Any "measuring factory" for the track output from a reasonable used accelerator gives enough work to at least a 704-type computer. Programming practices of the PANG-KICK type were assuming a computing capacity of that order, and a decision was made at CERN, in December 1959, to enable CERN - by hiring a to follow a similar programming policy. During the first half of 1960, programming work at Berkeley became influenced by the prospect of getting a 709 in the autumn of this year. The two essential additional facilities offered by the 709 are a far larger fast memory and the possibility of input-output in parallel with the actual computation. Programmes making use of these facilities are no longer compatible with 704 (but essentially compatible with those written for a 7090). The wish that CERN should remain in touch with programming work done elsewhere was one of the main reasons for accepting the IBM offer, made in August, of a 709 instead of a 704 (at a slightly higher hiring price, but since all hiring prices were considerably reduced at the same time, the newly agreed price for the 709 remains substantially below that agreed previously for the 704). b) Some of the principles, at present under consideration in the U.S.A. with a view to increase the efficiency of scanning and measuring by a substantial factor, may involve the use of computers larger than 709 or 7090, and faster than The cost of such a computer (purchase price of the order of $ 10 million) might seem not unreasonable when compared to the total cost of the millions of events it might have to handle, but it seems obvious that developments on this scale should not be envisaged in Europe on any reasonably foreseeable time-scale. We propose therefore to base our processing developments on the use of computers comparable in size to 709 and 7090 (the additional advantage of speed offered by 7090 and, presumably, by some other transistorized makes will become necessary after some time, when the maximum computing time obtainable with the 709 becomes insufficient). 3. POST-FRANCKENSTEIN DEVELOPMENTS The absolute necessity of putting considerable effort in the development of new processing technique is clearly recognized in America and action is under way. We would like to take out of the discussion here all developments which require many years of work, or the use of computers which are much larger than the ones envisaged for CERN. This excludes any device which automatically searches for events. The search will have to be done by man for the time being.

4 Page 3 We therefore only discuss developments which mechanize and thus speed up and simplify the measuring of events. Such devices are especially important for the measuring of simple events with high statistics. a) The Reaper Since Berkeley appears to be particularly interested in high statistics experiments, and since the Reaper development has been pursued there for a certain time, at a substantial expense, it was recently decided that Berkeley should concentrate all of its new-development effort on this device, although its somewhat limited scope is often recognized. b) The Flying-Spot Digitizer (Iep-X) The practical feasibility of this device was first demonstrated by Hough and Powell (CERN, 1960) and their communication attracted considerable attention at the Berkeley Conference (September, 1960). The immediate promise of the device, in its mechanical form developed at CERN, is similar to that of the Reaper (faster measurement), but the principle seems to be capable of a gradual deepening, leading to a completely automatized searching and measuring. One of the first possible improvements consists in using a cathode-ray scanner instead of a mechanical one; the most advanced models of CRTubes seem to have recently approached the required accuracy. Very recently (October), M.I.T. showed an interest in this line of approach, for which it is uniquely qualified; once electronic scanning is introduced, the development may become oriented towards the use of specialized computers which are familiar to M.I.T. but little known in other laboratories. 4. DEVELOPMENT AT CERN Firstly, the Iep Group at CERN is engaged in the work of manufacturing machines of the Franckenstein-Iep type. After completion of this construction work, in 1961, the effort will be put on the improvement and adjustment tasks mentioned above. Secondly, we are faced with the essentially new work of operating and maintaining a "measuring factory". This implies carrying out the actual measuring work on the material which will come in from the different bubble chambers. It also implies the organizing of the work at the seven Iep machines in the most efficient way, the stream-lining of all processes involved and the development of an effective book-keeping organization.

5 Page 4 Thirdly, efforts will be put on the continuation of the Iep-X development at CERN. The Hough-Powell device will be developed into a prototype whose properties and capabilities will be studied, and its programming developed, etc. The necessary contact with American development will be maintained by sending people over there, for shorter or longer periods, and by accepting people from America at CERN. It is obvious that the establishing and maintaining of a "measuring factory" together with the continuation of the development work on Iep-X involves additional people, so that the sum total of necessary manpower must show an increase in 1961 compared to The staff of the Iep Group will be increased, but in view of the general limitation of manpower in CERN, allocations for the "factory" and for the desirable continuation of the Iep-X development are to be mutually balanced. However, it seems unwise to allocate all available power to the "factory" and to stop altogether the continuation of the Iep-X development because of our restriction in manpower. It would seem wiser to keep new developments going, even if the data evaluation is somewhat cut down. There are two reasons for this: a) we must be prepared for the future, when the present apparatus will become obsolete; b) we must try to keep the most talented people, who are the ones interested mainly in the new developments. It is therefore planned to allocate roughly 6 to 8 staff members to the Iep-X development who might be joined by a few visitors from abroad. It is expected that the picture evaluation at CERN, together with the processing done at Universities in Member States, will remain at a satisfactory level.