Joi Ito's Web

Joi Ito's conversation with the living web.

I'm going to quote David's whole post because it has a bunch of good links.

David Weinberger
Metadata without tears

Peter Merholz, AKA peterme, has an excellent article at Adaptive Path called Metadata for the Masses:

But what if we could somehow peek inside our users’ thought processes to figure out how they view the world? One way to do that is through ethnoclassification [1] — how people classify and categorize the world around them.

He takes and Flickr as examples of "ethnoclassification" (a phrase he tracks back to Susan Leigh Star),. (I am enamored of the branch of ethnoclassification on exhibit at if only because people have started calling it "folksonomy.") He looks at the benefits. Then he addresses the problems, and suggests the paths out of the forest we're making for ourselves.

Jay Fienberg points us also to Jon Udell's article on "collaborative knowledge gardening." I've also been looking at some related issues (e.g., here, here, here, here and here), but Peter has the advantage of knowing what he's talking about.

I totally agree that this "ethnoclassification" is really an amazing solution to the metadata problem. Although, as they point out, there are some problems, I think that we'll find solutions. I feeling very taggy these days. I think there should be more cross-site tag linking. Blog categories, wiki pages, music meta data, and many other things can be "tagged". TAGCON 2005! Sorry. Just kidding.


I should rewrite ideas of this post with the light of today's services :)

Wired magazine did an excellent write up on the current problem of metadata as it relates to photographs. I've posted my comments and a link to the article at:

It’s encouraging to see that at least one web-aware VC is aware of the huge issue of page and site descriptive tagging.

Personally, I have given up hope more or less that any industry standard will be effective in making online files such as web pages comprehensible to crawler-based technology. Even if definitional consensus is achieved, the primary problems remain: 1) persuading publishers to tag billions of old pages along with all new ones, and 2) policing their honesty (since a bot can only do so much.)

In designing the platform underneath out prototype next-generation search site, we placed special emphasis on better ways for a website owner to describe content and functionality to the search tool’s indexing and matching algorithm. Our proprietary method allows us to do a much better job of ensuring relevance, as well as helping the site owner get much more (legitimate) exposure to search users.

More importantly, our method addresses all the types of descriptive records that are *not* presently constituted as web pages. This is, we think, the real breakthrough, since it means the “tagging” scheme can encompass *all* types of information. This is what searchers would really like to have: a tool that brings together in a coherent search results page all the sites and resources that are truly relevant to her search - - including the *commercial* and event info, too, at her option, of course.

3 TrackBacks

Listed below are links to blogs that reference this entry: Get down and get taggy.

TrackBack URL for this entry:

Given my recent experiences - nearly all positive - with metadata enabling services such as and Flickr, it's no surprise that I'm seeking ways to expand metadata hooks into new areas of my overall information management scheme. Joi's recent... Read More

Interesting; on the heels of my post yesterday about filing vs. tagging comes Joi Ito, pointing toward a post by David Weinberger that seems to resonate: Peter Merholz, AKA peterme, has an excellent article at Adaptive Path called Metadata... Read More

a DLA client - DLAs in action Living Design Case Study - - Flickr's gone and done it again - they keep updating and not standing stilll - good lesson for others..... nCube: The End of a Fantasy - tracking these folks really used to matter to... Read More