Re: Java and HTML and well known socket numbers

Paul Prescod (
Tue, 4 Jun 1996 12:36:34 -0400

Message-Id: <>
Date: Tue, 4 Jun 1996 12:36:34 -0400
To: James Aylett <>
From: Paul Prescod <>
Subject: Re: Java and HTML and well known socket numbers
Cc: MegaZone <>, Russell Holt <>,

>And Paul Prescod writes:
>> Well, if you have a big data source (let's say the complete works of
>> Shakespeare) and I want to do a really wonky query (let's say it involves
>> linguistics based interpretation of the text), I might just upload my applet
>> and you can bill me for the CPU time it takes up.

At 03:12 PM 6/4/96 +0100, James Aylett responds:

>Yes, but isn't this a usage in which a daemon is a bad way of doing
>things? Bearing in mind that you're charging people, it makes more sense
>to send the applet to a person than a computer, in which case no new port
>is needed because SMTP works quite fine ...

Why involve the human. The "server" can accept the demon, execute the query,
debit the user's e-cash account and upload the result.

>In addition, and even if you weren't charging people, how could you stop
>people say uploading a Java applet (or application, more likely) where it
>just spent a while chugging away on your server doing part of an
>encryption breaking job then communicating the result back. Would you
>really open your server to such abuse by running a daemon?

Presumably the Java execution environment would be sufficiently restricted
that the only meaningful thing you can do is work on the provided data (or
database). Sure, you could completely ignore the data and work on something
else, but as long as you are paying your $.00001 per CPU cycle, I don't
care. In fact, if I have a sufficently fast Java-machine, people might just
upload random computational programs without using _any_ resources on my
computers (except my CPU and ram).

 Paul Prescod