CHITAB – a “poor man’s” shortcut to computer processing of linguistic data
Jussi Salmela, Viljo Kohonen · DSpace repository (University of Tartu) · 1977
CHIIAB -a "poor nan's" shortcut to computer processing of linguistic data 1.Background.The primary purpose of the computer programme CHIIAB is modest: we hope it to be of use mainly to individual linguists or small projects with limited resources, who want to cope with sizable corpuses involving delicate classifications.In such a case processing the data manually will soon become laborious.To be more adaptable to linguis tic data processing, the present programme introduces some improvements over similar programmes already existing in various programme libraries (e.g., in HYLPS, in the Univac 1108 system of the University of Helsinki, and in the SPSS, for Dec 10).These improvements include the possibility for alpha-numeric coding, the use of up to ten "filter variables" to ex tract from the corpus precisely the desired variables for cross-tabulation, and the possibility to pick, out of the classifications of any variable, only the frequencies of any individual classificatory principles {e.g., codes 1,3,8,A,C, out of the total range from 1-F).These improvements mean a more economic use of the card space, and a more versatile use of the com puter.The basic idea in the system is that, instead of processing both the text and the coded symbols, only the symbols are fed into the computer.This solution naturally excludes certain kinds of research, such as vocabu lary frequency studies, but it is adequate for frequency counts and cross tabulations of, e.g., various semantic, syntactic and textual features.It is thus adaptable for a variety of research purposes.An important advan tage of the system is that it is remarkably cheap: the punching of the cards is fairly quick and straightforward, one card can accommodate a large num ber of classifications, and the computer processing is also quick.For the benefit of the individual researcher, who is frequently un sophisticated in computer technology, we have a l s o attempted to make the actual use of the computer as simple as possible.Thus, in order to have the computer carry out the desired cross-tabulations, the user only needs 82 CHITAB -a "poor man's" shortcut to computer processing of linguistic data Jussi Salmela, Viljo Kohonen Proceedings of NODALIDA 1977, pages 82-86