Subject:      Re: Web neurons
From:         john@eco.powernet.co.uk (John Middlemas)
Date:         1996/05/26
Message-Id:   <4o9vak$cgs@nuntius.u-net.net>
References:   <4nrc1e$dnk@power2.powernet.co.uk> 
	      <31A33F34.3886520C@redhat.com>
Organization: U-NET limited
Reply-To:     john@eco.powernet.co.uk
Newsgroups:   comp.infosystems.www.authoring.html



Elliot Lee wrote:

>So what do you want us to do about it?

Initially, just to give feedback, like you're doing. 

>Doesn't look like it has any provisions for processed data to be
>returned to the initiator.

There are several ways this can happen. The target can (after
processing) fire back INPUT(S) to the initiator in the same way the
initiator did to the target. Also, the target may not fire anything
back to the initiator but it may fire off other Web neurons that do.
Even way down the line the target could could get info back. Who knows
this could even be days later? The user would come later to one of his
Web pages and may find changes have occured (security could be built
in, I think).

Email (or "Webmail") could be transmitted this way. This could now
include graphics etc direct to chosen pages.  Also, direct editing
instead of uploading to your service provider!

Other  HTML extensions probably required::

<CODE>
Put at top of HTML to tell the target server, logic flow extensions
such as FIRE etc are present. It must then interpret code sections.

<NO_RETURN>
Do not return page to initiator. Perhaps it is a code page only.

To further answer your question, if say the target code actually
changed the target page and there was not a <NO_RETURN> then the
modified (and therefore processed) target page would be returned to
initiator. I don't know if this would be useful?

I like the word "initiator".

>I think HTML is meant to be a formatting language rather than a
>programming language - that's what makes it so popular.

Ordinary Web browsing would not be affected.

>Sounds like a decent idea, but I can imagine a lot of security problems.
>For example, with this, one might be able to create a distributed
>password cracker

>with computing power beyond imagination, and half the computer systems
>of the world would fall very quickly. 

I don't know much about security but I don't think it should deter
progress. Each Web neuron page should contain a full list of all its
initiators. If an outsider tried to call it could be detected on the
spot. To be even safer,  URL's in the FIRE command could be encrypted.

If everyone was nice we wouldn't need to worry. Why don't you all just
be nice!

Thank you very much for your interest. From the other replies I have
been getting I think there are some possibilities with this idea.
Sorry to take so long to reply, it has been a busy week.

---------------------------------------
John Middlemas