The existence of high speed computers with their large capacity memories makes feasible the concept of generating complete sets of derived data in the problem of spectrum classification. This possibility leads to a closer examination of the relative merits of the three types of generating operations as judged according to the probability of generating real data and according to the generated statistical noise level. Examination shows that the treatment of the observed data in terms of the concept of Zeeman tags is by far the most fruitful approach to be taken. A quantitative evaluation of the derived data is also given.
A specific set of computer routines designed to do both the computational and bookkeeping operations of the Zeeman method on the Oracle is described.
© 1959 Optical Society of AmericaFull Article | PDF Article
OSA Recommended Articles
J. R. McNally and P. M. Griffin
J. Opt. Soc. Am. 49(2) 162-166 (1959)
P. W. Ford
J. Opt. Soc. Am. 49(9) 875-877 (1959)
Barton J. Howell
J. Opt. Soc. Am. 49(10) 1012-1021 (1959)