Are we mushrooms?

[josep...@ucwphilly.rr.com]

Google Sets has been on the air for just a month shy of 3 years. Many
potential users (including me) have been frustrated trying to
understand what Google Sets aims to accomplish; others love it,
whether they understand its underpinnings or not. As far as I know,
Google Sets has not responded to the queries and puzzles sent to it.

This reminds me of an "ancient" Philadelphia or Pennsylvania saying,
quoted to me many years ago by a former co-worker, frustrated by his
work conditions:

"We're like mushrooms. We live in the dark and all we get fed is
sh-t."

Either Google Sets has rules that govern its results which it MUST
reveal to potential users or it has no RATIONAL rules, in which case
Google ought to pack it in and discontinue the project.

We, beta-testers, are being used, I'm afraid, by Google as free
slave-labor for a product that has never been seriously defined. It
may be fun for some, but it's no longer fun for me and I do not intend
to return to this site. I do hope, however, that Google management
will manage to acknowledge to us what they intend to do.

JoB

[via...@tiscali.it]

why so much
I think it is afirst direction into sematic web.
they must have some web ontology and look up for similar items. I am
not sure how did they get the ontologies from. (it doesnt just work
with common names, but also with.. soccer team names!)
it's nice to see that some of the latest research in AI come to life,
eventually. Even if you don't find a proper application right now.

Anyway I think it must have some use. Imagine you're writing an essay
and want to find related concept to a given one: maybe a novelist that
have difficulties in a description ?? a scientist that have to recall
some terminology quick??
answer to questions like which footbal teams? or simply wants to recall
names of famous brands out of few that you know before going shopping..

anyway you're not forced to use this tools..

pavel*

[josep...@ucwphilly.rr.com]

Pavel:

I would like to think, as you do, that "SETS" has "must have some use."
What is that use? Have you yourself used it for the purposes you
enumerate: "Imagine you're writing an essay and want to find related

concept to a given one: maybe a novelist that have difficulties in a
description ?? a scientist that have to recall some terminology

quick??" I sometimes write long essays, although I'm not a novelist.
I am a scientist who sometimes needs to "recall
some terminology quick??" I am NOT a person who "simply wants to
recall
names of famous brands."

Of course, I'm not forced to use the tool. However, since I'm
interested, like many others, in questions of Artificial Intelligence
(if there is such a defined discipline), I think it is perfectly
rational of me to ask what does GOOGLE(!) itself thinks this tool
should be used for. Not only this, but to answer some persistent
questions about what the semantics, grammar, ontologies and so on that
are assumed in the construction of this engine. I admit that some
users have been amused, entertained, perhaps enlightened, using GOOGLE
SETS. I cannot remember that anyone has claimed that having used
GOOGLE SETS has resulted in results such as a great poem or short
story, a new mathematical theorem, a novel experiment or theory in the
sciences. Perhaps I'm wrong, but I'd like to see what these results.

In any case, I see no excuse for Google not responding to the many
queries about what GOOGLE thinks that Google Sets might be good for. I
hope that this finishes my comments on this topics.

JoB

[eri...@gmail.com]

> > >
> > > We, beta-testers, are being used, I'm afraid, by Google as free
> > > slave-labor for a product that has never been seriously defined.

Why does it matter how Google operates, or towards what ends?

Our own evolution had no roadmap, and did not require constant
supervision.

I think letting us (the masses) beta-test these programs is essential
for Google's growth, in a sort of Rodney Brooks way.

How else could Google learn except from other people?

[josep...@ucwphilly.rr.com]

You write "How else could Google learn except from other people?" What
evidence do you have that Google learns from other people? In the
"good old days" not so long ago, I was able to communicate directly
with some of the programmers at Google to clear up some problems. This
appears to have completely stopped. I believe that Google is now in
the business of maximizing its advertising revenue rather than trying
to introduce really novel ideas. That may be an excellent business
strategy, but it is not one that encourages "other people to be heard."

That is to say that Google is probably a business like any other: it
seeks to to maximize its bottom line, not really to satisfy its users.
It that suits you, good for you. It doesn't suit me.

JoB

[josep...@ucwphilly.rr.com]

Another thought: if you try (more correctly, if I try) to send
feedback to Google Sets, the feedback function is disabled. How is
Google learning from us?

JoB

[kic...@gmail.com]

I find google sets very useful when I'm booking travel. I'll type in a
few of the discount travel websites that I know and Google sets will
spit back several more places where I can look for cheaper flights or
hotels. Also I was in the market for a suit and I wanted to know what
kinds of brands were availalbe and I used google sets. I think its
already a fairly useful product you just have to get creative with it.

[nate.kel...@gmail.com]

as much as i hate to denounce anything that google does i am also
struggling to find the real worl applications of google sets and i
support the idea of google letting us know what it intends to do with
google sets.

[nate.kel...@gmail.com]

as much as i hate to denounce anything that google does i am also
struggling to find the real worl applications of google sets and i
support the idea of google letting us know what it intends to do with
google sets.
josep...@ucwphilly.rr.com wrote:

[kulu42]

im not as concerned with uses as everyone else here. what i would like
to see is a format like google sets, except that one could potentially
place five random words or phrases in the query, and then see what
sorts of relationships exist between these.

ambulance, sweater, hunting bears, dogma, transformers

the most interesting google experiment i have enjoyed, is the
googlisms, which can often be random, undefined, nonsensical
characters.

folistmongo is :vava

i would stop trying to ask google what your purpose is with google and
use your imagination. or maybe you could do a google search: why do i
fear living laboratories?

[nick.r...@gmail.com]

I am currently doing a dissertation for my masters on interactive query
expansion (IQE) and the terms that Google Sets return seem useful in
this context.

Many search engines use query expansion behind the scenes in order to
return more relevant results. For example, if you enter the search term
"jaguar", you may also be interested in lions, tigers and other big
cats. With automatic query expansion, the search engine will append
terms to the query such as "lion" and "tiger".

One problem when searching the web for information is that often there
is a mismatch between the user's query terms and the terms used by a
website author. For example, a user may be searching for "fridge"
whereas a website may use the term "Refrigerator". Using IQE
techniques, the user of the search engine could append search terms to
their original query in order to retrieve more relevant results. This
is similar to how people search the web now. They enter a search term,
see the initial results and either add search terms to their existing
query or alter it entirely. This process is known as relevance
feedback.

I hope I haven't bored you :-)

Nick