home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
icon
/
newsgrp
/
group96b.txt
/
000065_icon-group-sender _Tue Dec 10 22:36:26 1996.msg
< prev
next >
Wrap
Internet Message Format
|
1997-01-02
|
3KB
Received: by cheltenham.cs.arizona.edu; Wed, 11 Dec 1996 10:27:15 MST
Date: Tue, 10 Dec 1996 22:36:26 -0600
Message-Id: <199612110436.WAA32077@ns1.computek.net>
Mime-Version: 1.0
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
From: gep2@computek.net
Subject: Icon vs. Java
To: icon-group@cs.arizona.edu
X-Mailer: SPRY Mail Version: 04.00.06.17
Errors-To: icon-group-errors@cs.arizona.edu
>Now, with the Internet, interpreter systems are back in the
name of portability of object code.
Trendy, yes... fad, maybe. I'm not convinced (at this point anyhow) that this
is a lasting thing... most of the Java stuff I've seen is strictly gratuitous
and can be readily dispensed with altogether.
>Suddenly the Intel CPU is not the only game in town,
It never has been, in theory. Remember SPARC? Alpha? HP-PA? And others?
None of those went anywhere major, and PowerPC isn't likely to, either. The
other companies would love to create a market for their chips via these silly
"Network Computer" things, but I predict those will fizzle in the marketplace in
much the same way that the Philips Interactive CD did, and IBM's Micro-Channel
did. In the case of the Network Computer, there simply isn't enough of a price
advantage to offset all the incompatibility disadvantages and limited feature
set.
>...and code has to run--unchanged--on multiple CPU architectures all over the
world.
The simple fact of the matter is that it DOESN'T need to. 90-95% of the market
is Wintel PCs, and (as demonstrated by all the Web sites which don't work right
on anything other than Netscape) a lot of people are perfectly content to see
their stuff work on only a subset of the client machines, as long as that's a
big enough and valuable enough subset of the total.
>Recompiling is not the answer; the Java runtime interpreter, which combines
abstractions of both the computer and the network, is seen as the solution.
Java is hardly the first such "machine-independent interpreter"... Pascal P-code
would work just as well, as would even Databus/Datashare for that matter.
There's nothing terribly amazing about Java.
If I were particularly interested in Java-like features, obviously I'd much
prefer to write in ICON rather than Java. But in fact I think the point is
largely moot since I think it makes more sense not to do this kind of thing at
all... use the Web for what it's designed for (i.e. distribution of information
CONTENT) rather than for what it's not (i.e. program loading).
I realize this isn't cowtowing to all the latest fads, but I've been in this
business long enough to see a *lot* of basically useless fads come by (and go,
most of them).
Gordon Peterson
http://www.computek.net/public/gep2/