[3789] in java-interest
Re: Fundamental flaw in Java library distribution scheme
daemon@ATHENA.MIT.EDU (Bill Joy)
Tue Nov 28 01:22:53 1995
Date: Mon, 27 Nov 1995 21:03:56 +0700
From: wnj@sw.Central.Sun.COM (Bill Joy)
To: wnj@sw.Central.Sun.COM, david@longview.com
Cc: java-interest@java.Eng.Sun.COM, flar@bendenweyr.Eng.Sun.COM
>Think about the impact of even small problems in the language design in
>5 years. If I was Sun I would not put *anything* into Java that was
>not incredibly well understood, and vetted by the academic community.
to be honest, we have spent the last 5 years having arguments like
this, and are very aware of, e.g. hoare's classic paper on language
design, which says roughly this.
every feature has been proposed to be in or out many times;
we have reached the point of exhaustion... someday we'll write
a book reproducing parts of the discussion.
in particular, we discussed taking out method overloading.
i was in favor of this, but it judged
to be very useful for the simple cases,
namely t.print etc.
saying t.printint, t.printreal, ... makes programs less
readable not more, and readability is a major goal of the language.
for the more complicated cases we made the (compile-time) rule
compatible with the more sophisticated rule used by languages
like dylan/cecil. we did this because it seemed to reflect the
negative experience in the lisp and self communities from more complicated
dispatch (e.g. prioritized, etc.) dave ungar (who suffered through
this in the self world) helped us to understand the issues and i believe
we "got it right".
in the end, as actually practised there seems to be no problem
with overloading. what you describe as a "fundamental
flaw" is probably more accurately a "curiosity".
wnj
-
This message was sent to the java-interest mailing list
Info: send 'help' to java-interest-request@java.sun.com
