Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sdd.hp.com!hp-pcd!hplsla!billh From: billh@hplsla.HP.COM (Bill Harris) Newsgroups: comp.edu Subject: Re: Language Use Message-ID: <11210001@hplsla.HP.COM> Date: 18 Mar 91 20:01:58 GMT References: <1150@ra.MsState.Edu> Organization: HP Lake Stevens, WA Lines: 42 I just found an article by Curtis, Sheppard, Kruesi-Bailey, Bailey, and Boehm-Davis entitled "Experimental Evaluation of Software Documentation Formats", published in "The Journal of Systems and Software", vol. 9, pp. 167-207, 1989. Briefly, they studied the effectiveness of certain types of software problem-solving behavior as a function of documentation formats and styles. On page 198-199, they describe how they removed the effect of programmer variability from the results of their study. They found that the number of years of professional programming experience did not correlate to performance on any of the 4 experiments they ran. On the other hand, the number of languages the programmer had used did correlate positively with success on 2 of the 4, and it was on the borderline of significance on a third. They claim this is consistent with reports in a prior study ("Modern Coding Practices and Programmer Performance" by Sheppard, Curtis, Milliman, and Love in "Computer", vol. 12, no. 12, pp. 41-49, 1979), which showed that the number of years of programming performance didn't correlate with performance, but breadth of programming experience (number of programming languages known, etc.) did. (I have not yet read this article.) Incidentally, the subjects (programmers) in the 1989 report were all professional Fortran programmers with an average of over 5 years professional experience. My conclusion is that you should pick the concepts which you wish to teach and then select a language which is effective at teaching those concepts and meets other constraints which you may have, as one of the other posters mentioned. It would only seem beneficial for students to be exposed to a variety of languages over their careers at the university. Having hired software engineers in the past, I can say that a person's understandings of the fundamentals was more important than whether or not they already knew the language we were using at the time. Bill Harris billh%hplsla@hplabs.hp.com Hewlett-Packard Co. Lake Stevens Instrument Division ms/230 8600 Soper Hill Road Everett, WA 98205-1298