shebs%defun.utah.edu.uucp@utah-cs.UUCP (Stanley T. Shebs) (02/01/88)
In article <6950006@hpfclp.HP.COM> fritz@hpfclp.HP.COM (Gary Fritz) writes: >> [...] The hardware vendors fell into the trap of assuming >> that good software is easy to build - a common misconception of hardware >> types who "once wrote a 200-line Fortran program". In fact, hardware vendors >> have *rarely* succeeded in producing good software of *any* sort, [...] > >This is an unfair statement. This relates more to software engineering in general than to Lisp systems, but I think it reflects reality. Time and again, we see the hardware first and then the system software tacked on as an afterthought, then the vendors are surprised and unhappy when customers avoid their proprietary but hastily written code in favor of something that had a little more effort put into it. Upper management is probably the most to blame - if they have an engineering background, it's more likely to be in hardware, and the prejudices just seem to be inevitable. ("CS? No, we find they're too undisciplined. We prefer to retrain people with other degrees, say Math or EE.") >HP Common Lisp was developed by a group >of extremely qualified software engineers, many with extensive previous >experience in building Lisps. Granted, the implementation had some >problems, but those were due more to historical/etc. reasons than to >"hardware types who once wrote a 200 line FORTRAN program". After over two years of using HP CL, I'm familiar with its characteristics. I'd like to know why these "extremely qualified software engineers" didn't stand up and demand a thoroughgoing rewrite, instead of patching and patching PSL innards (imagine my cynical lack of surprise to find a "PSL" package, and a "system-lisp" flag that disables type checking indiscriminately). I'd like to know why so much effort was put into a flashy but inferior editor, instead of into the nuts and bolts of a decent compiler. I'd like to know why Spice Lisp wasn't credited as the source of much of the runtime system (as evidenced by certain quirks and bugs that should have been cleared up before HP CL was sold for big bucks per copy). At least to me, calling oneself a "software engineer" implies a certain level of professional responsibility, including whistle-blowing if "historical/etc reasons" are resulting in a bad product. At Utah we were (and are) grateful for massive donations of HP equipment, but our attempts to return the favor by suggesting improvements to early versions of HP CL were met with complete silence on HP's part. Again, I suspect it was mostly upper management's fault, but did the software engineers involved ever question HP CL's schedule or staffing or requirements? stan shebs shebs@cs.utah.edu