Thoughts, links, and essays from Peter Merholz
petermescellany   petermemes


Archives before June 13, 2001

RSS Feed

Adaptive Path (my company!)

About peterme

Most of the Time
Oakland, CA

American history around the time of the Revolution, figuring out how to marry top-down task-based information architecture processes with bottom-up document-based ones, finding a good dentist in San Francisco Oakland
Designing the user experience (interaction design, information architecture, user research, etc.), cognitive science, ice cream, films and film theory, girls, commuter bicycling, coffee, travel, theoretical physics for laypeople, single malt scotch, fresh salmon nigiri, hanging out, comics formalism, applied complexity theory, Krispy Kreme donuts.

Click to see where I wander.

Wish list
Show me you love me by
buying me things.

Track updates of this page with Spyonit. Clickee here.

[Editor's note: began as a site of self-published essays, a la Stating The Obvious. This evolved (or devolved) towards link lists and shorter thoughtpieces. These essays are getting a tad old, but have some good ideas.]
Reader Favorites
Interface Design Recommended Reading List
Whose "My" Is It Anyway?
Frames: Information Vs. Application

Interface Design
Web Development
Movie Reviews

Lou Rosenfeld Needs Comments Postings. Posted on 12/11/2001.

And permalinks, while we're at it. See, in his latest Bloug post, he talks about the power of the "humble hyperlink." Go over there and read it. Then come back.



Okay. So, you've read it, right? Well, what Lou is basically describing is how Google works, if I'm understanding him, and understand Google.

Google pays attention to not only the connections between pages, but what words are used to click on them to get there. This is why Calamondin is the first Google result for the query Judith, even though her name never appears in the text of her site (it does appear on a mailto link, but I doubt that's why she ranks so highly).

And Lou's request for a magic "MORE" button is akin to Google's "Similar Pages."

In fact, Google having figured out what Lou figured out is part of the reason why they're a magic search engine.

11 comments so far. Add a comment.

Previous entry: "one thought on information architecture."
Next entry: "Vanity Time Travel."


Thanks for reminding me that Bloug doesn't support comments and permalinks. It will in a matter of days, I'm promised. "Manana, manana!..."

But comments aside: while simliar, this is NOT how Google works. Unless I'm really missing something.

Google certainly uses a pattern matching algorithm to establish a retrieval set, and a link-based popularity algorithm to rank those results.

I'm suggesting something that I believe is a little different:

* The page where you START from provides a certain type of context that I don't believe Google takes advantage of. In fact, I don't think Google does anything with establishing context linguistically.

* The page you END on after clicking a link can also provide some good context *if* you like it. Many search engines already will get you "more like these". But I think there may be interesting things to be learned by an engine during that transition from START to END. Maybe.

* I don't know of a tool that allows you to submit a LINK without reentering it, but I imagine this is technically quite trivial to implement. Does anyone know of tools that convert link labels into search queries?

Anyway, that's my reaction. Of course, Peter remains partly responsible for this idea, whether he likes it or not...
Posted by Lou Rosenfeld @ 12/11/2001 08:22 AM PST [link to this comment]

It could be based on Google's "related:" search option, which is one of the few things that Dave's Quick Search Deskbar doesn't do yet.

Here's a Google search looking for pages related to and I think that gives pretty good results, in fact, darn good results, with almost no oddballs in the list. You have to build the query manually, but it's pretty close to "show me stuff related to THIS site and THAT one", which is nearly the same as "show me stuff related to where I AM and where I'm GOING"
Posted by Andrew @ 12/11/2001 09:29 AM PST [link to this comment]

I almost have a bookmarklet working that does what we're talking about above. Clicking it rewrites all the links on a page into Google "related sites" searchs using both the current URL and the URL of the clicked link.

However, it seems like Google doesn't actually allow multiple parameters in these queries; it just uses the first one.

Anyway, you can see the bookmarklet sort of work at If anyone sees how to get this to work, let me know!
Posted by Andrew @ 12/12/2001 06:29 AM PST [link to this comment]

"One word, Benjamin... Alexa"
Posted by matt @ 12/12/2001 06:44 AM PST [link to this comment]

I don't think Alexa does exactly the same thing as descibed above, which is to sort of triangulate related sites using BOTH the context of the source page and the context of the destination page.
Posted by Andrew @ 12/12/2001 08:00 AM PST [link to this comment]

sorry andrew - i'm obviously not getting it... one of you guys is going to have to draw me a diagram... ;-)
Posted by matt @ 12/12/2001 10:16 AM PST [link to this comment]

Well, I suspect that in Lou's dream I was probably grinding that same old axe of mine about the importance of linguistic context in IA.

It sounds to me like what Lou wants is a cross between Google and Northern Light's notorious Little Blue Folders. Such a system would use the Northern Light algorithm to identify the collections the START page and the END page belong to; then the Google algorithm would kick in to refine the result set within (the intersection of?) those collections.
Posted by jjg @ 12/12/2001 12:21 PM PST [link to this comment]

Matt, I know just what the diagram would look like! Start with three circles intersecting....
Posted by Andrew @ 12/13/2001 06:14 AM PST [link to this comment]

Andrew, how'd you guess?
Posted by Lou @ 12/13/2001 12:49 PM PST [link to this comment]

Google calls themselves "fair". There is nothing "fair" about being on page 6 or 7 for our relevent keywords when some sites with absolutely no relevance at all are making it on page 1. Can anyone explain this to me?
Posted by Adam Medeiros @ 01/11/2002 05:13 PM PST [link to this comment]

That's nothing Adam. Until last month, Google didn't list Amazon on the first page of results for a search on "Books."
Posted by Jeff @ 04/17/2002 11:52 AM PST [link to this comment]

Add A New Comment:


E-Mail (optional)

Homepage (optional)

Comments Now with a bigger box for text entry! Whee!

All contents of are © 1998 - 2002 Peter Merholz.