[OT] We need a language
Brian T Rice
Fri May 30 14:38:01 2003
On Fri, 30 May 2003, PB wrote:
> "We need a language that lets us scribble and smudge and smear,
> not a language where you have to sit with a teacup of types
> balanced on your knee and make polite conversation with a
> strict old aunt of a compiler."
> Paul Graham, "Hackers and painters".
Well, this certainly got a lot of responses. It has well-thought-out
responses on LL1 as Fare mentioned. The stuff here is silly.
May I re-interpret "we need a language" as a call for HLL definition?
There are certainly a lot of misconceptions among today's posters.
Certainly the designs in http://tunes.org/HLL/architecture.html should
explain that we're not talking about a single language, but something more
akin to a network. The spec work at http://tunes.org/~water/spec/ points
to a similar explanation, and also explains that the primary point is to
be able to define subsets of functionality and such that make those
optional proofs that Jeff mentioned possible. The spec defines a subset as
any situation where something is restricted in /any/ way at all, which
includes typing of attributes or configurations. This is both necessary
for a useful system and for a safe and high-performance system. How else
can you assume that something won't happen except by not letting it happen
or observing that it doesn't (noting that real-time observation violates
the principle anyway)?
Slate certainly doesn't attempt to do any of those things, and Max doesn't
either, really. They're projects to demonstrate and exercise the use of
various aspects, like persistence, interfaces, run-time architecture, etc.
Do any of these things not make sense to people, or are you content to
reply about something which has nothing to do with TUNES with ramblings
that ignore what TUNES has to teach?
Brian T. Rice
LOGOS Research and Development