Tom Chi  

The Epistemology of Search

November 5th, 2004 by Tom Chi :: see related comic

This comic actually happened to me several years ago. The only difference was that my coworkers and I never solved the problem. Instead of applying the empirical approach, we used web-enabled devices, phoned friends that had net connectivity, and tapped our somewhat rusty recollections of physics — all to no avail.

Sometimes it’s the fancy pants technology that keeps us from finding the simple answers. Technology can hide as much as it reveals and a lot of hidden assumptions lurk in those bits and bytes. The network now surrounds us, redefines what is known and what is knowable. When we search we feel like we are connecting with so much data, that an empty search means that perhaps the subject itself is null. Conversely, a search returning a rich result set is probably important, and the first few hits are the very most important. These are just a few of the assumptions that underlie our new world of knowledge. But let’s take another look –Searching for “Tom Chi”, I find that I largely dominate the first page of results. But does this imply that I am the most important Tom Chi? From my own research into this very important topic, I must confess that I am not. Tom Chi’s range from orthopedic surgeons to kung fu masters, and i’d be hard pressed to claim superiority to a kung fu master. What it does mean is that I am the most technologically connected of the Tom Chi’s, and also that I’m probably online too much.

Similarly, all things that we search return not so much ‘importance’ as ‘connectedness’ within the limited domain of web publishers. To give an example — I don’t know much about scimitars (traditional south east asian ceremonial blades), but if I were to put up a page about metalworking for scimitars, in a couple months it would probably be a top search result in the domain. It is not because I am any sort of expert, but rather because I am relatively more expert than most web page publishers. Still, I’ve never held a scimitar or even seen one outside of a museum.

KC calls this “the world according to Google” (or WAG for short), a seemingly complete space populated with domain experts and pundits. The algorithms approximate importance and time relevance, but ultimately boil down to something more akin to ‘fame’. Madonna is not an expert on the Kabbalah, but when she speaks about it, many people tend to listen. Similarly, Dan Cederholm can write about slippers and get a ton of comments. His post is already the 4th result for a search on ’slippers’, beating out slipper manufacturers who have been in the game for 100 years.

Such is the state of the network today. In many ways we already trust in its ability to find answers to our problems, but it has quite a ways to go before it truly embodies what we’ve already come to expect.

15 Responses to “The Epistemology of Search”
Gabriel wrote:

I don’t see what’s wrong with that.
I liked the concept in the comic. too much information and tools keep us from getting the work done… but in the comic example, the work is to know how long the coin takes to get to the floor or the work is to have fun finding out? I think that both got the work done with the right tools, but they just had different goals.

Now, with the google relevance in the article, i truly think this is just another layer of information google gives us.
If i’m looking for a kung fu master, your name will not appear on the first ones if i use the keyword “kung fu” along with the name. That’s ok.
But what about the extra info i just talked about? Well, if i’m looking for you, a Tom Chi that i’m aware that has a strong online presence ,aka don’t get out much :) , i know that i can only type Tom Chi and find you on the first results.

PS: what’s that post preview thingy? put your money where your mounth is :P

Sarah Brodwall wrote:

> If i’m looking for a kung fu master, your name
> will not appear on the first ones if i use the
> keyword “kung fu” along with the name.

It might now. :

noah wrote:

Heh… I see this all the time!

Technology, here as world-wide networking, allows maximizers to be even stronger maximizers than they ever could have been before–it’s about getting the BEST answer, not just the appropriate one for the moment. Whereas sufficers are just as happy to drop the coin, count, and accept the result. (Check out “The Paradox of Choice” by Barry Schwartz. It’s a great–and relatively light–read on the matter.)

Google is purely democratic on results, which as we all know, does not mean accuracy of relevance or content. Obviously, the problem is when people overlook this factor, which is why I usually distrust any report that factors the number of hits on Google on a subject as a representation of anything besides how many links are pointing to that text pattern on the Internet. The answer is to overlay this kind of activity with metadata, like an extension of XFN: marking relationships *and* overall trust of the link itself. Such a browser could then display what your approximate relevancy to a new document or trust level of a new source might be based how your “inner circles” rank it.

Engnu wrote:

One of the dangers Tom pointed out is the “seemingly complete(ness)…” of Googlespace. It may in fact be the latest manifestation of symbol replacing experience as a way of knowing. ie: Once we are accustomed to the abstraction “tree”, do we then lose all living contact with real trees?

The antidote of course is not to reach for the google search button as an ultimate authority on everything, and to know that there is an immeasurably vast real world outside. This is admittedly a hard thing to do for the technologically-baised set given that with the right connectivity it is so much easier to google something than to use other means that may be at hand to find out about something.

A case in point: My boss recently asked me to research glass block pavers (just like the kind you see on sidewalks:) He suggested that I might do well talking to So-n-so who is a local authority on the subject having used just such a product on a recent project. So what did I do instead? You guessed it - I deftly executed a series of Google searches, and in a matter of minutes found a promising supplier of glass block pavers, complete with architectural details! This does not mean that I would not have found a superior solution if I had picked So-n-so’s brain on the matter, but it goes to illustrate how seductively convenient this technology really is!

Allan Rojas wrote:

Yeah, it is true that the first organic results might not be the most important; but how will the future generations know where to find the real important ones?

Tom has a great point here, get someone to search for ’slippers’ in 2 years and the organic results might throw tens of tweaked blogs talking about the subject…

‘The World According to Google’ is something that should be treated now… before it’s too late. I mean, how do you measure the quality of the information that’s put in front of you, if you can’t see other sources to compare?

Gabriel Mihalache wrote:

+1 for using “Epistemology” in non-philosophical conversation. You rock!

Tom Chi wrote:

Well there are several ways to attack the problem. Google could continue to expand the space of what it covers (like their recent foray into print) to eventually capture knowledge from unwired communities. Although this still doesn’t address the issue of communicating what Google knows about the space and what it doesn’t. It also doesn’t create a distinction between what is topical and current vs. what is richly valuable, but not current.

Dan’s post about slippers is topical and current, but it is not a rich collection of resources on slippers. Depending on what the searcher’s intentions are, they might either, one, or both. To include this data would make for an interesting info design problem… hm.

Jitendra Jain wrote:

The Site for keyword has moved to number 3, thanks to thousands of blogs. we decide what remains on Top and what doesnot.

Jez wrote:

And there’s me, having to look up the word Epistemology!

Dmitry Nekrasovski wrote:

Regarding the slippers issue: The Googlespace creates the problem, but it also offers a
solution. :)

Engnu wrote:

Filters, filters, filters! As an end user of information technology, I would prefer to have the means to do the filtering myself, and not to cede control to the whims of some AI context/word parser.

Paulo Eduardo Neves wrote:

The best example is still searching for “Apple” in Google. All the results of the first page are about the computer company, nothing about the fruit or the english record company.

Gabriel wrote:

Yesterday i thought about this post… i was looking for some buildings the architets from SITE did for a department store called BEST (now extinct) in the usa. So i typed without thinking into google “SITE BEST building” then just before hiting return, i realized what would happen…

Engnu wrote:

I know what you mean. Checking out your example, it took me several tries to figure out that “best products” along with “site architecture” yielded halfway decent search results. Failed searches frequently offer clues on what not to type in, but overall it can seem like a perverse game of 20 questions to find what you need.

Nick Douglas wrote:

Gabriel: That’s why, when thinking up project names, I use Google-uniqueness as a criterion. I used to run a blog called Popageorgio, a reference to “National Lampoon’s Vegas Vacation,” because there were no online references to the name spelled that way (though there are many Papageorgios).

The now-defunct site is still the lucky result, because my friends haven’t changed their links.

Leave a Reply

OK/Cancel is a comic strip collaboration co-written and co-illustrated by Kevin Cheng and Tom Chi. Our subject matter focuses on interfaces, good and bad and the people behind the industry of building interfaces - usability specialists, interaction designers, human-computer interaction (HCI) experts, industrial designers, etc. (Who Links Here) ?