Java �ber Alles?
From: Tom Christiansen <
[email protected]>
Subject: Java �ber Alles?
Newsgroups: comp.infosystems.www.authoring.cgi, comp.lang.perl.misc,
comp.lang.misc
Message-ID: <
[email protected]>
References: 1, 2, 3, 4
----------------------------------------------------------------------------
In comp.infosystems.www.authoring.cgi,
[email protected] (J.M. Ivler) writes:
... and we all know that Java will kill both of the languages [tcl
and Perl] soon enough (as useful pieces of WWW).
Oh we do, do we? I find your assertion more than merely somewhat suspect.
Here's why:
Three broad areas of concern vis-�-vis Java arise: its complexity,
propriety, and security. I will today consider in detail only the first
matter. But as just one example of the others, though, realize that despite
all possible assurances of its alleged pristine and wondrous security, there
shall remain many, many, many sites that simply but absolutely shall never
ever permit running Java viral code downloaded to their system from
elsewhere in the world. Just go check on this at any of our national labs or
military installations if you don't believe me on this one. Period, finis,
and QED.
Setting aside matters of security and propriety for the nonce, let's examine
the first of Java's three critical issues: complexity.
While I have admittedly not plumbed Java's most profound depths, and
notwithstanding the fact that Java certainly has some interesting lessons to
teach us, it is to all appearances yet another bondage-and-discipline,
ivory-tower programming language that will be forever out of reach of the
``casual'' programmer. By casual programmer, I do not mean those cerebral
priests in their secluded cloisters dedicating their entire lives to writing
the next great operating systems, newsreaders, command shells, distributed
network admin tools, number crunching spreadsheets, WYSIWIG editors, or
flashy new programming languages. No; instead I'm talking about that
incredibly vast majority of simple users out there who are only trying to
get their jobs done, jobs that require a mere modicum of practical
programming rather than stellar science.
Java seems a lot like an interpreted, somewhat kinder-and-gentler version of
C++. Lovely! It is thus destined to raise fierce passions on both sides
based on that aspect alone. But being kinder and gentler than C++ doesn't
make it kind or gentle per se: while Purgatory may be less unpleasant than
Hell, it's still not a nice place for a holiday. Likewise, even though Java
may be only a bare one-third of C++'s complexity, that may not suffice:
one-third of a big number is still a big number. Java's saddled with a great
deal of B&D complexity stemming from its C++ ancestry and the perceived need
to stress formalistic language concept ideals over the quotidian concerns of
normal end users.
Perhaps you might care to explain to a simple user why his objects get
``finalized'' at non-deterministic times rather than when he's done with the
object? Do you want to explain to that simple user why from function B, he
can't call function C if C has an exception in its signature that B isn't
explicitly dealing with, even though he called B from A, and A does handle
C's exceptions? I know that I for one most certainly would not relish such a
wretched task.
Unless it should change dramatically, I cannot see how Java can ever become
the Language of Choice for part-time, non-MSCS, casual users who have a wee
bit of programming to do (read: almost everyone who does any programming
anywhere). I will concede that some avante guard software engineers and
mathemagicians may latch on to this shiny new Rubik's programming bauble,
but I don't find Java sufficiently ``mundane'' for those outside the
esteemed Ivory Tower of cyberlinguistics.
Scheme is another great example of an interesting language beyond of reach
of mortal men, despite the fact that Scheme is a fine little language of
subtle power which every CS major should study. But it's not accessible, so
it loses. What's to stop Java from falling into the same black hole? Lavish
wealth squandered on marketing blitzen to get Java stuck on the front page
of the New York Times' Science and Technology section or the San Jose
Mercury News every few months? We've all seen that kind of push before:
Pascal, NeWS, OpenLook, even Novell's ephemeral acquisition of Unix. Et ubi
sunt?
In summary: I believe that until and unless Java should become substantially
easier for normal people to use, your prediction of tcl's and Perl's
impending javanese eclipse is as unlikely an event as that familiar but
fallacious ``imminent death of the Net'' doomsday cry. By extension, tools
and programming languages like Python, BASIC (particularly the ubiquitous
Visual BASIC environment), REXX, and even the lowly Bourne shell and its
still less respectable brethren aren't likely to be on their way out the
door even for WWW applications any time Real Soon Now [TM] .
So sorry, no pictures at 11. [:-)]
--tom
----------------------------------------------------------------------------
See also Felix Sebastian Gallo's rejoinder to my message.
