ILL - Final comment
Initially neutron scattering employed state of the art numerical control to provide 24/7 operations for precise mechanical diffractometers and spectrometers to maximise use of the feeble intensities available. By the time the ILL was created paper tape and electromechanical comparators evolved into small control computer systems individually crafted to perform the same functions. Time of flight techniques developed from event recording on magnetic tape, through multi-channel analysers, to using small control computers.
A tour de force was achieved in 1975 at the ILL in measuring the inelastic scattering from 3He on IN5, which required more than ten days counting without re-zeroing the counts. A few years later the spectrum could be seen after an hour on IN6, which still only had detectors covering 3% of the solid angle.
As detector systems evolved (q.v. the "Second souffle") data volumes increased. Neutron studies still used important features of the absolute nature of the intensity measurements to perform fundamental experiments. Each year the Vercors group tried to fund an additional 1 square metre of detectors for IN5.
Spare computing capacity and general availability of Fortran on control computers lead to scientists gaining the ability to exam data immediately, and perform better experiments, especially where it was necessary to extract the science from multiple measurements. In diffraction the completed sequences of scans were analysed on the central computer.
In the mid 80s the advent of 32-bit computing on instruments offered a wider choice for distributed computing, ending with ethernet connections around the Insttitut.
Data treatment was initially performed on even more expensive main-frame computers. As graphics evolved the value of hardcopy output from the original storage scope Tektronix terminal was more and more appreciated. Raster graphics became possible as memory prices dropped and microprocessor speeds increased. Colour graphics using domestic microcomputers became a reality when these used colour lookup chips to generate the colour raster; 35mm polaroid slides were used for hardcopy.
The supporting staff for these activities diminished. Machine code and assembler programming disappeared from instrument control, these tasks being performed by libraries which later added necessary networking features in the current era. For the central service, the operators have been relieved of tape loading and physically moving data now that many terabytes of storage are available on-line in a shared database. Many scientific programs have been introduced from external laboratories and universities. Overall the residual programming activity at the ILL needs little mathematical support.
In contrast assistance is required and generously offered to support the very large park of personal computers, including Macintosh, Microsoft and Linux systems, together with the network infrastructure. Fortunately the hardware reliability of most of these single-board systems is very high.
Because of the impact of computers on every aspect of the ILL the total investment in hardware and manpower for computers, networks, and control electronics is still a major item of expenditure. Data treatment is less rigorous; there are greater volumes, and the scientific result can often be shown simply presenting pictorial evidence of effects of external scans of temperature etc, on more or less raw data. Absolute cross-section measurements, e.g. isotope variation for liquid structure work, is rarer, and fewer such experiments are attempted than in the past. Hence the value of the graphical packages included with the commercial software (IGOR, Matlab etc), though a considerable effort is still spent on adapting the GUIs for daily use by visitors.
Today digital control is everywhere, at low cost, using networked microprocessors which can easily handle parking a car,and even driving it safely. The solution to any computer related problem, from mobile phone to personal computer is well known to everyone; a reboot will, hopefully, restore normal operations!
Ron Ghosh, Epsom, November 2017