Re: About the question what a robots is (was Re: PS)

Rob Turk (rturk@austin.ibm.com)
Tue, 16 Jul 1996 11:38:25 -0500


Michael Schlindwein wrote:
> > >Great idea, but again is begs the question, "What qualifies as a robot?" For
> > >example, does WebCompass (which retrieves two levels deep to summarize
> > >findings) count as a robot? What about other types of "user agent"
> > >interaction?

I agree. Defining "robots" with all these dogmatic, arbitrary
definitions is not going to really make things easier for web sysadmins
or webmasters, or robot developers. One hears all the time, or it used
to be heard, that web browsers are just as much agents as anything else,
and that their searching of bookmarks, hotlists, or user preferences is
also "automated" processes.

I'd like for as much as possible to be automatic while safeguarding 1)
the authenticity of the data 2) the security of the user agent and the
data sources 3) everyone's time. These things should be
optimizers...freers from the mundane...working smarter not harder...etc.
Really I'd like to have automated brokers that manage investments etc.
so well that eventually everyone can leave the offices and departments
of this world and spend their time researching, creating, teaching, or
doing whatever to improve life for all.

Will this happen overnight? No. People are going to write software
that -- Heaven forefend! -- may make unfair demands on machines, whether
by design or by accident, but by communicating about better ways to do
things the software will improve. The functionality of agents will
improve, etc. I'm not interested in having the entire Net mirrored onto
my Mity Mite Hard Drive every other night, or getting indexes of
people's private stuff. I respect privacy above almost everything,
especially on people's web sites.

I think that a great homepage would be one who could spin off users into
various directions by the author(s) based on user agent validation and
encryption based security for private files, password based security on
directories and applications, and for the public... a "automated agent"
friendly "what's new!" type document standard that would be updated as
the results of author(s) activity on the server (which would include
some sort of integrated style-sheets or document template gateway) and
it would all be magically timely and interesting to many potential
audiences willing to shill out nanobucks for the info...or maybe macro
bucks as well...depends on the commodity, I suppose. If someone could
guarantee that a document would either entertain the hell out of me or
make me money I'd be willing to pay for it, or authorize my user agent
to make it available to me at my leisure...but how would that be
guaranteed?

I suppose that I'd like the discussion to focus more on applications
than bitching about etiquette, when it seems like it would be unfair to
say that a person is "wrong" in a matter of etiquette when that person
can't be "gently but firmly" reminded of the existence of a FAQ
somewhere about all the stuff that's been thrashed out during these
discussions. Everyone is entitled to their opinions, their own
interpretation of what is "proper".

-- 
Rob Turk <mailto:rturk@austin.ibm.com> Unofficially Speaking.
About the time we think we can make ends meet, somebody moves the
ends.
		-- Herbert Hoover