PROCEEDINGS OF A SYMPO - IAEA Nuclear Data Services

4 HOREN and WEINBERG
these two separate threads from earlier works: Scientific Information, in
this instance, Nuclear Information, and Criteria for Scientific Choice.
F o r just as science itself, and certainly nuclear science, must adjust to
limited resou rces, so scientific compilation, and in particular compilation
of nuclear data, is sim ilarly constrained. Thus the problem of priorities —
what to do first, where to allocate resou rces in science — also faces the
community of com pilers. They too must decide what to com pile and what
to leave for later. Can we establish a-p riori criteria of choice for scientific
compilation? Can such criteria be regarded as m ore than a philosophic
exercise?
THE NECESSITY FOR CHOICE
The com pilers of scientific data have in recent years lagged behind
the producers of scientific data, and have had to establish p rio ritie s. In
nuclear science, technical developments have greatly magnified this discrepancy
between the m ass of data and the com p iler's capacity to handle it.
The use of high-resolution solid-state detectors, as well as other
im proved techniques, has multiplied the number of recorded bound states
per nucleus five- to ten-fold during the past decade; and the widespread
utilization of automatic processing has created a heavy glut of undigested
nuclear data. There are now som e 3000-4000 nuclear-structure scientists
who publish about 3500 papers annually. The data contained in each paper
varies anywhere from a single number — e .g . the reporting of a half-life
measurem ent, or spin, etc. — to thousands of numbers. F o r example,
Mühlbauer has observed m ore than 2000 gam m a-ray transitions below
1. 4 MeV following neutron capture in 152Eu [ 5] ! Ignoring the quantities —
level energies and properties — that can be deduced from this particular
data, but, considering only the energies, intensities, and uncertainties
on each quantity, one has to deal with 4 X 2000 or 8000 numbers!
The history of the Nuclear Data P roject from 1959-1972 illustrates
how this increase in data has com plicated the life of the com piler of m a ss-
chain data. P rior to 1963 the number of compilations per m an-year was
3. 5-5, with each compilation having 100 times the quantity of data reported
in an average research paper; today this number has fallen to about 1. 8
compilations per m an-year. Since 1959 the number of research papers
per m an-year (at least in the United States) has remained relatively constant
at about one. The average quantity of-reported data, however, has increased
by a factor of ten, and the ratio of data per compilation versus that per
average research paper has remained about constant at 100/1. Thus the
com piler has increased his ability to handle data by a factor of four to
five — i. e. (1. 8/5) X 10 — but this has not been sufficient to cope with the
ten-fold increased production of data.1 Thus one can make a serious
argument for relatively m ore, rather than less, money going into secondary
treatment of data.
1 W e h a v e trie d to e s tim a te th e c o s t o f g a th e r in g and c o m p ilin g n u c le a r d a ta . T h e a v e r a g e research
p a p e r rep orts a p p r o x im a te ly 30 m e a su re d n u m bers (in c lu d in g u n c e r ta in tie s ), w h ic h is e q u iv a le n t to a b ou t
$ 15 00 p er n u m b e r; in th e U n ite d S tates an a v e r a g e re s e a r ch p a p e r co s ts a p p r o x im a te ly $ 45 0 0 0 . D ata p rod u cers
a re g e n e r a tin g s o m e 110 0 00 n u m b ers a n n u a lly . A n a v e r a g e D a ta P r o je c t c o m p ila t io n c o n ta in s a p p r o x im a te ly
3 00 0 n u m b ers a t a c o s t o f a p p r o x im a te ly $10 p er n u m b e r, o r o n e p e r c e n t o f th e p r o d u c tio n c o s t.
In this c o n n e c t io n th e authors w o u ld li k e to p o in t to th e im p o r ta n c e o f e s ta b lis h in g a u n ifo rm and in te r n a tio n a l
sy stem o f k e y w ord s and in d e x in g as a m ea n s o f a m p lify in g th e c o m p ile r s ' in c r e a s in g ly d iffic u lt jo b o f k e e p in g
up w ith th e d a ta .