Thou shalt have no other gods before the ANSI C standard 1594

Tom Linden

If you mean elaborate in terms of your PL-I point, I'm not sure. I'll try, but don't take my opinions too seriously; I've never written a program, not even "Hello World", in PL-I.

I know that PL-I was a terrifically successful language as measured by the success of software written in it. On the other hand, it is also commonly cited as a *bad* example of programming language design, for including features it would be better off without, simply because some customer requested them (including its name; "New Programming Language" collided with the abbreviation of a customer: the National Physics Lab). I'll not enter that debate.

I'd describe the bottom-up approach as "given what we have, how can we implement what people need?" The top-down approach is "here's what people need, how can implement it given what we have?" I know of definitions in the literature that buttign more mathematically specific meanings to the terms, but that's how I'd describe their general use.

The PL-I design approach of surveying users for their feature requirements is neither, specifically, top-down nor bottom-up. PL-I, like other languages, separated implementation detail from specification, but specified only features with known implementations. It largely built on features in existing languages, which is mostly a bottom-up approach.

Contrast that with APL. Iverson initially devised APL as an improvement upon existing mathematical notation, not existing programming languages.

-- --Bryan