From: Tony Finch (firstname.lastname@example.org)
Date: Wed Sep 27 2000 - 13:14:34 PDT
>4. Will software/hardware be able to eventually pass the Turing
>Test, or will this require a biotech approach?
>7. No software language has survived indefinitely, other than
>1's and 0's. Will there eventually be one that does, much like
>English seems to be doing?
Lisp, Fortran, and Cobol seem to be doing a good job of surviving
>9. Is it possible to create a language and IDE which rejects any
>attempts to insert defects, and is productive?
This is where we get to mention the Entscheidungsproblem and thereby
sound impressively learned. However, although the general case of this
problem is insoluble, various special cases are soluble, and indeed
optimising compilers rely on this fact. However, sohpisticated
compilers that can spot tricky problems at compile time rely on
mathematical theories about the programming language they are
compiling. Since most languages in use today have a specification like
a dogs dinner, compilers for those languages have a fairly limited
understanding of the code they are compiling (and frequently any
possibility of useful understanding is scuppered by the
specification). State-of-the-art compile time analysis and bug
spotting relies on languages with precise definitions and
sophisticated type systems, which tend to discourage less
mathematically-inclined programmers. Development environments for
these languages tend to focus on writing compilers for these languages
and therefore lack practical features that you might expect.
>10. Will it become possible to prove any system has no defects?
>14. Will the three rules of solid (what's the proper term?)
>programming - sequence, decision and loop - be replaced by a
>whole new approach, perhaps allowing a whole new language
>generation? What is it?
Functional programming and logic programming have been around for a
-- en oeccget g mtcaa f.a.n.finch v spdlkishrhtewe y email@example.com eatp o v eiti i d. firstname.lastname@example.org
This archive was generated by hypermail 2b29 : Wed Sep 27 2000 - 13:19:32 PDT