Where should the type information be 169
Herman Rubin operators,
Herman:
your part of To be clear on my part, my problem is that he typically is NOT telling us what he needs as a...
I have been foollowing your comments on programming language on the internet for over ten years now. It is obvious that you have strong opinions, but, as in this case. I have great trouble understanding those opinions, and I believe I am far from alone in having that trouble. At least part of it appears to be due to a failure on your part to learn standard terminology, part of it may be a failure on your part to sit down and think systematicallt through your ideas, and part of it is a strong difference between your opinions and the opinions of others. In particular you often place a strong emphasis on syntax in your discussions, but from an implementor's point of view the most critical part of the language is its semantics, and much of your discussion of semantics comes across to me as superficial.
Yes, I understand that. TOPS-10 had oodles of customers who did exactly the same thing. The trick is to figure out what can be implemented that is not Rubin-specific but general-user...
In this case I don't understand your distinction between operators and functions. There are languages where the distinction has great meaning, APL and its descendents come to mind, but in most languages operators represent functions that are so widely used (or at least widely understood) that it is very convenient to have a shorthand notation for them. In such languages operators are often overloaded, but the resolution of overloading is done by semantic analysis, not by parsing. Further many such languages also allow overloading of functions, also performing resolution by semantic analysis. Compare X+Y with ATAN(X, Y). Finally, the distinction between compile time and interpretation is not usually a distinction explicit in language definitions. There are language design aspects that can make compilation easier, e.g., static typing and local scoping, but compilers have been implemented for a variety of languages that appeared initially to require interpreters.
Herman Rubin bringing Henry: 1. Rhetorical question. You have stated the equivalent of "I believe I could do it easily in what could be called a HLL buttembler" a large number of times over the...
APL is an exception in that many of its "operators" change the meaning of other operators or funtions, e.g. changing multiplication to matrix multiplication, and this form of operator can (at least sometimes) be resolved by parsing. However I have rarely seen you refer to APL in your discussions.
Where should the type information be 170
Alt Folklore Computers from Newsgroups