andrewt@basser.SUN (Andrew Taylor) (03/19/84)
I find some of the statements in Fernando Pereira's discourse on the "Realities of Prolog implementation" hard to swallow. Here are some of the more interesting. 1. It is rather difficult to write a good, elegant and space-efficient Prolog INTERPRETER in anything but Prolog, witness the fact that most Prolog INTERPRETERS are either VERY slow, space-extravagant, or have minimal debugging facilities. I have written (what I consider) a good, elegant and space-efficient prolog interpreter in C. It offers almost the same debugging (and other) facilities as DEC-10 Prolog. 3. C is a poor language to implement elegantly and efficiently interpreters for tagged languages like Prolog. This is what we in Australia call bullshit. C is an ideal language for writing a good, fast interpreter and indeed quite a few such interpreters have been written. I can not see any reason why prolog should be any different. Indeed, as I said above, I have proof of this. 2. Moral: if you want a good and elegant Prolog interpreter, write first a Prolog compiler and then write the interpreter in Prolog Unfortunately I don't believe that this is true. You may be able to produce an elegant prolog interpreter in prolog. However your interpreter will not be as fast or as space-efficient as a well written one in a language such as C. Hopefully this situation will eventually change. Fernando gives a detailed explanation of how and why C-Prolog was developed. I never disputed that for the effort involved the production of C-Prolog was a probably a good idea at the time. Its many users testify to this. But as Fernando says C-Prolog is only a stop-gap system. I think we need better software. This may involve much work but someone has to do it. Fernando seems to show disdain for "academic" elegance in software. I (and many others) believe that elegant coding and good software go hand in hand. I understood the great success of Unix was largely due to the elegance of its design and implementation. I don't think Ritchie & Thompson concerned themselves with "useful local maximums on the effort/benefit curve". Andrew Taylor ...decvax!mulga!andrewt:basser