David said:
Who said happiness had any part of it? Certainly your footnote did
not. You referred to 'dedicated' Ruby coders, actually - I was unaware
I was signing up to the new religion. I would also suggest that I am
one of those people who *does* lie in the intersection of the set of
language lawyers and hard-core hackers (if you will pardon my
assertion of semantic equivalence) - as my long-ish career in
programming on a wide variety of platform shows.
Hooray!! Another person on this list that's been programming a "longish"
time and is in the intersection of language lawyers (aka people who
appreciate the formalisms of computer science) and hard-core hackers!!
Have you been doing it long enough to have earned a living programming
in assembler?
Sorry for saying that the Emperor has a hole in his trousers, but I
really didn't say he was naked - because he is not. However, formal
specifications *are* hugely useful, assuming that you have learned how
to think rigorously enough to make use of them. And as the history of
some programming languages has shown, need not be a barrier to
innovation even within the language specified.
My take on the history of programming languages is that the hacks and
innovations have *led* the formal specifications, not followed them.
People had been doing concurrent programming long before CCS and CSP and
mobile programming long before the Pi-calculus, for example. In fact,
I'd go even farther and say that almost all of the innovations in what
we call "computer science" originated at the hardware level, with the
software being crafted to use the hardware and the computer science
being crafted to explain the software.
*cue the old geezer tape* ... on second thought let's not
My attitude is what I said it was. I'm having a fair bit of fun
programming in Ruby - enough that I'm actively thinking about ways to
implement a native-code compiler - but the way that the language is
specified is a problem to me from many angles. And the process of
discovering that specification by trial and error brings back to mind
the worst of the bad old days of cross-assembly development and hand-
patching device drivers use hex editors.
I can assure you, the "core" of the Ruby community, which I define as
the people implementing the interpreters and with active Ruby internals
projects, are aware that this needs to be done and soon. Other things
that need to be done and soon are performance tuning of the inner
interpreter, getting Unicode on the mainstream, and stabilizing the
syntax and semantics. Other people have pointed out the web sites where
most of this is going on.
I've briefly flirted with the idea of doing a Ruby implementation
myself, or, actually, resurrecting one called Carbone that's based on
gForth and vmgen. That's mostly because I'm a better Forth programmer
than I am a C programmer, and because I think "vmgen" is a masterpiece
of the intersection of computer science and hard-core hardware hacking.
But in the end, it looks like the Ruby 1.9 inner interpreter in C
will be as good or better, so I'm happy to work on applications and use
them to test 1.9 instead.
By all means, feel free to go ad hominem all you want - there's a long
and glorious Usenet tradition there. Many of us have felt deprived
since the departure of Erik Naggum from comp.lang.lisp, but the one
constant in the world is change...and crabby[1] hackers
david rush
http://cyber-rush.org/drr <- a very messy construction^Wwebsite
I was not aware of that -- I always thought falconry was cool, but never
bothered to dig into it.
