USERMILF@SFU.BITNET (Peter Milford) (08/05/88)
Hi, One issue not mentioned in the current discussion on the utility of FORTH, is the productivity of FORTH programmers. In C. Moore's paper on FORTH Astron. Astrophys 15 pg 497 (I don't have it in front of me, so the info is from memory) (perhaps dated now - does anyone have more recent references to productivity in FORTH?) Moore concluded that FORTH had several advantages for that project (instrumentation control): a) Increased productivity. The job was completed faster using FORTH compared to the alternative of FORTRAN/Assembler. b) Smaller computer. Because FORTH is a 'smaller' system, the required computer was smaller (less expensive). The net result was a completed job, with less manpower and lower costs. I have little doubt myself that FORTH is a more productive system for PRODUCING WORKING CODE. However other considerations, particularly of maintenance, etc. can easily lead one to the use of other languages. Why is FORTH more productive? Perhaps because the time spent on the edit-compile-link is greatly reduced? Perhaps because of the ease of testing individual 'words', making debugging simpler? Perhaps due to the extensibility of the language? High programmer productivity alone can be a good reason for switching to FORTH for many applications. Peter Milford. usermilf@sfu.bitnet or pmilford@solar.stanford.edu
jbn@glacier.STANFORD.EDU (John B. Nagle) (08/07/88)
I tend to be very skeptical of articles of the form "Why X is great" by
the author of X.
See "Experience with the programming language Modula" in Software
Practice and Experience circa 1978, by (guess who) N. Wirth. This was
Modula I he was talking about, which he now refers to as "an experiment".
You tend not to see articles along the lines of "How we wasted three
years trying to make X work".
Bear in mind that when Calvin Moore wrote that, the competition was
at the level of Tiny BASIC.
John Nagle