CHI-WEB Archives

ACM SIGCHI WWW Human Factors (Open Discussion)

CHI-WEB@LISTSERV.ACM.ORG

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Condense Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Sender:
"ACM SIGCHI WWW Human Factors (Open Discussion)" <[log in to unmask]>
X-To:
Date:
Thu, 10 Jul 1997 17:23:25 -0400
Reply-To:
Jeff Brandenburg <[log in to unmask]>
Subject:
From:
Jeff Brandenburg <[log in to unmask]>
Content-Transfer-Encoding:
7bit
In-Reply-To:
<[log in to unmask]> from "Michael DeBellis" at Jul 10, 97 01:26:16 pm
Content-Type:
text/plain; charset=US-ASCII
MIME-Version:
1.0
Parts/Attachments:
text/plain (50 lines)
Michael DeBellis wrote:
>
> If you say so. My point wasn't to get into an argument about weather
> forecasting. Just that to me it seems pretty dull and marginally valuable.
> It just bothers me that people think up interesting concepts such as agents
> and then can only think of such (at least to me) trivial examples of what
> they would do.

We all have our different obsessions, er, interests. :-)

> >So why should *I* have to wade through all 100% to get the 0.1% of good
> >content, if my agent can screen out 90% of the junk and only 50% of the
> >good stuff?  95% of what I wade through is still junk, but I've still
> >improved my odds.
>
> I agree with you there.  Although TV is limited to Buffy for quality there
> is a lot of good stuff on the web but its hard to find because there is so
> much more junk. I think part of this may be a question of different work
> styles. I would rather have a better search engine (and content that had
> more meta-information to help the engine determine if it was relevant) than
> an agent that I program and have run in the background.

I agree with your comment about styles.  Given my work and browsing styles,
I would rather have an agent that presents a small amount of filtered
information to me; if I have to browse for (or even search for) my target
among lots of tangentially related materials, I'm much more likely to be
drawn off on a tangent.  Naturally, I would *also* want access to the
better search engines.

There's a seductive and dangerous notion that filtering will *prevent* one
from seeing the critical bit of information one seeks, while search engines
and direct-manipulation interfaces will let you sift through *all* the
information to find the bits you want.  I'm beginning to realize that
flooding my attention with all the information I can scan, counting on
my focusing mechanisms to notice and sieze items of interest, just
doesn't work.  Perhaps it works for others.

> Which leads to one of my proposals for something I would like to see in the
> future: better standards to annotate sites and pages along predefined
> ontologies so that smarter engines (or agents) could find relevant stuff.
> Knowledge representation technology such as Loom and KL-One might be very
> relevant.

I'd be happy to see workable and usable ontologies.  I tend to be skeptical
about this goal, as you are skeptical about the AI technology agents may
require.  I'll be surprised if we can come up with a set of really good
standard ontologies -- but I *love* that kind of surprise!
--
        -jeffB (Jeff Brandenburg, Persimmon IT)

ATOM RSS1 RSS2