[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: What?



Carl Gay <carlgay@attbi.com> writes:

> I have to say that even though I think it would be fun to have
> those things in GD, what's much more important in my opinion
> is getting threading working, speeding up the compiler, and
> implementing libraries.

We certainly agree on these goals.  We discussed threading on the DHC,
but thought it could not be done safely without changing descriptor
representation.  Rob McLachlan has convinced us that it is possible to
do without, if some care is taken in the optimizer, but that was on
the mailing list a few days after.

Now speeding up the compiler was one of our main concerns.  After all,
we could get more work done if turnaround times weren't that bad.

Since a 10% speedup in the compiler would not be a big help, just
doing some optimizations wouldn't cut it.  We figured that having
interaction capability would be the big win needed to make work more
productive, and getting the other tasks at hand done faster.  An
interpreter for the intermediate code seemed to be the quickest route
towards that goal, and the best use of the limited time we could spend
together.

And more important than writing new libraries is making the existing
libraries work, there are after all about 600KLOC of Dylan code out
there.  So most of our work during the last two weeks was hunting
bugs.  Amongst new libraries that work are the Fun-O test libraries,
and the tests for Dylan and Common-Dylan.  As of today, the test
summary says:

libraries-test-suite summary:
  Ran 26 suites:  8 passed (30.8%), 18 failed, 0 not executed, 0 crashed
  Ran 67 tests:  56 passed (83.6%), 11 failed, 0 not executed, 0 crashed
  Ran 1987 checks:  1819 passed (91.6%), 87 failed, 0 not executed, 81 crashed

Anyways, at the moment bug fixing is our main concern.  We hope for
you people to write useful libraries, and frankly, the availability of
koala is quite a motivation for me.

Andreas

-- 
"In my eyes it is never a crime to steal knowledge. It is a good
theft. The pirate of knowledge is a good pirate."
                                                       (Michel Serres)