W3C home > Mailing lists > Public > www-talk@w3.org > May to June 1996

Re: Java and HTML and well known socket numbers

From: Matthew James Marnell <marnellm@portia.portia.com>
Date: Tue, 04 Jun 1996 17:13:09 -0400
Message-Id: <199606042113.RAA14120@portia.portia.com>
To: www-talk@w3.org
:>Once upon a time Matthew James Marnell shaped the electrons to say...
:>>Sorry, I missed this in the first reading.  Again, since nobody
:>>has asked this question, I will.  Why are we talking about running
:>>Java applets on servers?  Why did someone go through all that trouble
:>
:>Check on Sun's Java announcements.
:>
:>By 1997 they plan to have it so a client can assemble a Java applet on the
:>fly - say a search routine - and send that applet BACK to the server for
:>the server to execute or proxy to another server.  Little smart frames
:>running about to and fro.

Nice idea.  You remember the little dog in the AT&T commercial getting
his nice owner tickets to some concert.  Haven't seen it yet?  You will.
These little agents are nice ideas, and maybe we will see them by
97.  But does this mean that an agent looking for tickets will query
the java port to see if can run on Joe's server in Massifornia or will
it query all known ticket-[master|tron|server] ports and actually
run on those computers or will it just send a standard query to that
port, using that services API and run on the owners computer.

All these apps running to and fro, how can a java server on the
Java port know if should accept the app or send it onto another
java server?  Which other server?  I can see the person paying for
the execution, but how about all that bandwidth as it's trucking the
'Net looking for the right port to call at.  What if I set my server
to accept all apps, whether it can answer their question or not.
Hey, I need the money and whether they get anything meaningful back
from me, I'll still charge them for cycles.  Or I write a nice
distributed app, forge the auth and use it to split up the code
to break into some government computers, change a certain parameter
that was programmed into the app by me so that it doesn't have to
return anything to me.  Since I used the key card and terminal of
someone that had gone to the bathroom without logging out, all cycles
get charged to poor Joe and the FBI comes looking at your computers
and his for the breakin attempt.

The scenarios are endless, and this is assuming that we're not
just talking about web pages.

:>They made a big deal out of this during JavaOne - and several vendors of
:>HW and SW said they will be implementing it.

That's nice.  If I had a dime for every time something like this has
been said and later poo-poo-ed, I'd have my daughters 2 million dollar
college tuition for the year 2012 paid for, and that's just up until
now, not figuring for inflation and interest.  Chalk it up to "I'll
believe it when I see it, and I still don't see the relevance to
a Java well-known socket."

:>That's another thing - Several vendors are pushing Server Side Java applets,
:>or servlets, as an alternative to CGI for database interfaces, etc.

This is fine if you run Solaris, NT/95, the Mac, and probably AIX, AUX,
IRIX or DEC Unix.  Okay, I guess that's all that matters.  Or you could
use any number of the other things like python, PHP, etc that are still
more portable over more systems and already do this at any number of
sites on the 'net already.

:>There was a list of them discussed at JavaOne - Java is not just a web
:>thing anymore.  There is a lot of planning for clients to make up applets
:>and toss them to other clients, to servers, through proxies, etc - a bit
:>Gibsonian but effective.

"Don't believe the hype!"

Believe me I know that Java is more than the web.  That's what I'm
planning for, and that's what I'm working for.  But, Java is NT
(Not There) in any meaningful way, and won't be for some time to
come.  It's nice, it's cool, it's an egregious waste in 95% of all
places it's used at the moment, it's being squandered for nifty
graphics tricks, rather than used for other meaningful content
delivery and is nigh impossible to index the content it does
deliver at the the moment.

I like all the ideas for Java.  I'm excited about them.  With
Internet paranoia being what it is, Internet payment schemes
being what they aren't at the moment.  The state of encryption
export from the US being what it is at the moment.  The bad
press and FUD that outfits like OBT have laid on Java of late.
Any number of other things, like promised delivery dates for
key software and hardware components being missed, etc.  All
these things make me feel that not only are people putting
the cart before the horse (which hasn't really been born
yet in any meaningful sense) but are going about everything
half-assed in their excitement.

With much more hype, and the sluggish pace of delivering most of
it (relative to the Internet's way of thinking) the Internet could
end up being the most expensive burnout in history.  What did the
most recent "authoritative" Internet count say?  30 million?
Even it's 50 million and a lot more businesses than is assumed
now, we're still not over the meaningful hump yet.  If it all
doesn't start living up to the hype RSN, people are going to
start looking at Internet people like the boy who cried wolf,
and it could look really bad, not just for the WWW, Java and others,
but for the Internet as a whole.  I'm not trying to be a doom-sayer,
just trying to point out that Java doesn't need it's own socket
even when/if we're throwing apps around pell-mell.

Let's see, if I believed all the hype I wouldn't have Unix and
Macs, I'd be running NT and 95 or Taligent.  I wouldn't be doing Unix
consulting, I'd be doing NT consulting.  I wouldn't have a line to the
Internet, I'd have a line to MSN.  I'd be paying all my bills and
getting all my checks via the MSN.  I'd have a little agent getting 
me the best prices on CD's and hotdogs and having them delivered to
my house.  The computer industry would be completely ruled by Wintel.
Yadda, yadda, yadda.  The moral of the story, don't put all your eggs
in one snake.

Matt

PS, this is all I have to say on this matter further, aren't you happy?
Received on Tuesday, 4 June 1996 17:13:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 October 2010 18:14:19 GMT