[1099] in java-interest
Re: overloading of operators
daemon@ATHENA.MIT.EDU (Walter Smith)
Fri Aug 18 17:35:08 1995
Date: Fri, 18 Aug 1995 10:32:01 -0800
To: Philip R Brenan <phil@bga.com>
From: wrs@newton.apple.com (Walter Smith)
Cc: java-interest@java.sun.com
>What is the point of Java? Its most unique point, byte code portability,
>will surely be adopted by C++ compiler as soon as Java proves whether it
>is worthwhile.
You should switch to caffeinated coffee! :-)
First, bytecode portability is not the only interesting thing about Java.
(In fact, I want an unportable native-code Java compiler.) Anyway,
Microsoft has been compiling C to bytecodes for years, so by your argument,
the disaster already happened.
The point is that C++ has been flown into the ground, and we need a better
general-purpose language. The fastest way to popularize such a language is
to attach it to something else that's popular--in this case, the Internet.
(The meme-virus potential of the Internet is orders of magnitude greater
than Unix's was, and look what happened to C!)
C++ continues to grow even as it approaches standardization. Most of the
people who like it actually like some earlier edition of it. Once it's set
in ANSI stone as the insanely complicated thing it is, and people have to
deal with it all, I predict a backlash.
I know a _lot_ of C++ programmers. I have never heard anyone say they
actually liked it. They just put up with it because there's nothing
better.
Most of the complexity of C++ is there for the sake of "performance". It's
really kind of a noble experiment. Once people realize what you get when
you sacrifice everything for some abstraction of "performance"--measured in
processor cycles for small expressions--I think they'll wake up and realize
it's not the number-one priority C++ assumes it to be. Performance is only
important for 5-10% of your program. Why should we have to screw up the
whole thing to make that 5-10% go fast?
C++ will go the way of all languages. It will be popular with some
forever. There will be plenty of legacy code written in it. But something
else has to take over. Given the imbalance between software and hardware
technology that exists today, I think the pendulum has to swing back
towards functionality rather than raw speed. (It could hardly swing
further over!)
In fact, it's already started. Look at Visual Basic, for example--not a
great language, but the right idea. IBM is pushing Smalltalk. OpenStep is
ramping up in popularity, and will carry the Objective-C virus with it.
The GNU folks are working on Scheme for the Unix masses (surely a bastion
of C++dom). A miracle might happen: people might discover Modula-3. (And
of course, I must modestly point out NewtonScript.)
At the least, I expect more two-tier programming (something nice for the
90% and C++ for the fast 10%). But in that case, who needs C++? Just use
C.
Cheer up. Things are more interesting than you think.
- W
---------------------------------------------------------------------------
Walter Smith Internet: wrs@apple.com
Newton Software AppleLink: WALTER.SMITH
Apple Computer, Inc. PGP key at ftp://ftp.apple.com/pub/wrs/PGP-key.asc
+1 408 974 5892 E1 20 C3 0A DE 27 89 06 0B 35 08 65 0C FB A7 41
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com
