[Sigia-l] the future of search

Thu Jul 25 15:56:27 EDT 2002

In a previous life, I spent much time thinking about web searching for
IE. Here's some thoughts I remember having:

There are many different kinds of desired experiences that people use
search to satisfy. 

1) Search as a specific chore: If the goal is truly to find a single
specific resource ("what is the population of Uganda?"), then the best
search service is one that takes the shortest amount of interaction
time. The psychic search idea, where magically the system knows what it
is I really need, and delivers it, would truly be ideal. Short of magic,
the goal for designers is to find ways to minimize the chore of
searching, and improve the ease and speed at which people can
communicate what they want, review potential choices, and make use of
the information they need. 

But often even when people know precisely what they want to find, they
have trouble articulating it to something or someone else. Ambiguity
reigns in language. Even worse, even if the customer's task is specific,
their may be multiple sources of data, that conflict, and a single
answer is not desireable. (word definitions, historic accounts, etc.).
So the chore has lots of sub chores that a designer has to support.

2) Search as a fuzzy chore: Often people have only a vague idea of what
they want. They have some problem they are trying to solve, but they may
have difficultly identifying or deciding on precisely what approach they
should take. (My sink is leaking. Is there an available local plumber,
or should I find a how-to manual?). The process of searching in these
situations also becomes a process of alternative exploration. The result
lists from a mediocre query often provides several different potential
alternative approaches to solving the problem, and this can be
beneficial. The customer might have never been exposed to those
potential alternatives if they were delivered a single specific mythical
best answer. So often a list of good results is better than a single
great result. 

3) Search as serendipity: Search can be a means of pure exploration.
Some people search google for people's names. They search for friends
from high school, or put in words for topics they like just to see what
they might find. Some people enjoy the hypertext experience, as an end
unto itself, and like to use search as a way to amplify their powers to
do so. Sometimes it's the least relevant result in a search results page
that is the most interesting or valuable one to receive. 

4) Search as research: For the knowledge worker or project manager,
search can be part of an ongoing process of research. To prepare for
writing a non-fiction book, or examining the prospect of aquiring
another company, search is just one part of a complex set of
inter-related tasks. The concepts of annotations, referencing, searching
and bookmarking all come together. To support someone using search in
this context might require more complex tools, and perhaps a different
learning curve. The depth of interaction can be much deeper than "search
as chore". 

So when I think about the future of searching, I think about systems and
interactions that try to capture the very different goals and processes
that people have when they search. Perhaps there are patterns of
behavior that can be mapped to heuristics that capture which search
experience I'm trying to create for myself. Or learning from amazon, and
ideas of influenced wandering, search could know more about me, and my
tendancies to focus or explore, and adjust the releveancy of results
based on a wide set of factors.  There may be certain semantic
relationships between topics I tend to explore on (fuzzy/serendipity),
and topics I tend to want specific answers on (search as chore). Somehow
this intelligence has to work conservatively - I'd rather - with mininal
false positives. The stronger it's influence, the greater the risks are
when it misinterprets or mispredicts my behavior 

I think from an IA/Usability perspective, we tend to fixate of search as
a specific chore, which is a limted view of the search experience
(though it's one that easier to examine). Beyond my little framework
above, I'm sure there are other ways to frame the spectrum of search
experiences - which implies an even greater range of potential designs
and models for what search experiences should be like. As always, the
design challenge is how to unify the divesity of needs and approaches
people want and need, within the sanity of a single set of design and
engieering decisions. 

-Scott

Scott Berkun
Design & Usability Training Manager
Microsoft Corporation

-----Original Message-----
From: Joe 10 [mailto:joe at joe10.com] 
Sent: Thursday, July 25, 2002 10:59 AM
To: christina wodtke; sigia-l at mail.asis.org
Subject: Re: [Sigia-l] the future of search

I think it will be when we're all happy with what we have and can all 
give up on all this searching.

Honestly though, I believe it was Clement who said something to the 
effect of "in the future there won't be 500 channels, there will be 1 
channel, and it will be my channel", or some such thing. Maybe it 
wasn't even him...

Developing some sort of "Preference Map" based on pre-filtering 
content through a continually evolving set of personalized heuristics 
will certainly cut down the noise.

If I want to ask for a reference on refinishing hardwood floors, I 
don't want to ask Google and I don't want to ask the SigIA list, but 
I'd probably ask my neat-freak friend in Sausalito who's persnickety 
about his floors. Having a predetermined set of channels you've 
personally set up because they return the results *you* want and can 
rely on will inevitably cut down search time.

Then, if there's a way to map the parts of my "preference map" 
against the portions of my social networks "preference map" which I 
find helpful (such as "tune my shopping agent into CW's wine choices, 
but tune out her sports interests" for instance) with continually 
increasing granularity then I'll spend less time searching at Google.

That is, my Father should not have to try to learn boolean logic. 
Just won't happen.

RSS and the ability to draw and aggregate content based on your 
preferences is moving us that way.

RDF and the Semantic Web initiative will help too. IA's really have 
to think of how to organize content not just so it's more human 
usable, but so it's more *machine* usable; both ends need to meet. AI 
isn't a bad thing, just the evil tin of IA.

Like I've said before, ( 
http://www.info-arch.org/lists/sigia-l/0204/0165.html ) let's not let 
the nerds shape the future of automated retrieval and categorization 
lest it look and feel thoroughly dismal.

Back to my meditative mood now,
/Joe

At 7:32 PM -0700 7/24/02, christina wodtke wrote:
>I used to think search was dull. But my "future of interface" brought 
>me in contact with kartoo, boolistic and other strange search engines 
>that showed me that search might have a long way to go yet. Information

>visualization, learning interfaces, semantic search... what's next? 
>I've heard google gives a talk where they say that someday a person 
>will just go stand in front of a computer and it will tell him the 
>answer to his search with out a query
>even-- which I think is I bit far off yet...
>
>Since we're all and a meditative mood anyhow, I thought I'd ask what 
>you think the next radical innovation in search will be.
>

-- 
Joe Tennis
Information Design Honcho
Joe 10 Enterprises
2430 5th Street
Studio L
Berkeley, CA 94710
510-649-1744
joe at joe10.com
http://www.joe10.com
Usability Consulting | Information Architecture | Interface Design for
Web, Wireless and Interactive Media
----------------------
Try the newer, gentler, standards compliant Joe10.com
------------
When replying, please *trim your post* as much as possible.

*Plain text, please; NO Attachments

ASIST Annual Meeting: http://www.asis.org/Conferences/AM02/index.html
ASIST SIG IA website: http://www.asis.org/SIG/SIGIA/index.html
_______________________________________________
Sigia-l mailing list -- post to: Sigia-l at asis.org
Changes to subscription: http://mail.asis.org/mailman/listinfo/sigia-l