I was talking my my friend Adam the other day about Wink, and he mentioned the “Wink Answers” feature. This is a typically ingenious Web 2.0 functionality mashup of search + wiki + tagging. Apparently, Wink is a tag-enhanced search engine — but once you search, you’re offered “Wink Answers” in addition to tag- and Google-powered search results. Wink Answers are simply wiki entries keyed to search strings. Given that a string like “mov video converter” is much more specific than any of the single words contained therein, this could be more useful than, say, the Wikipedia entry for video formats. Neat idea. Apparently, someone else has already taken this to its logical endpoint and mashed wiki + URL to create “Shadows” — clearly insane, but not surprising.

In any case, there isn’t yet a _Wink Answer_ for “mov video converter”. Given the results of my search yesterday, I now have the answer. The question is, why would I stuff it into Wink? I can blog it right here, Google (and a thousand other spiders) will crawl it, and the next time you search for “mov video converter” on Wink (or Google, or anywhere else), my post will show up, if it’s a good one. The only reason I’d shove my Answer into Wink is if I don’t have some other way to hang it out where spiders can see it — and if I’m on the net searching for info about “mov video converter”, then I almost certainly do.

The same problem/solution applies to reviews, which I’d guess are the broadest and deepest user-generated content out there. So why do so many people take so much time to write detailed reviews and post them in single, walled off databases like Amazon, Yelp, sfsurvey.com, etc.? It all comes down to the value exchange. You provide the content, the site provides the UI, and specialized economy of scale. The amount of personal content/data/facts that you’re willing to embed/give to a site is related to the utility that the site provides to you, personally, with regard to that information. But, more importantly, it’s directly proportional to the extent that particular site can act as a proxy for your desired audience.

As broad search UI’s become more powerful, the need for specialized UI’s to act as proxies to specialized audiences is reduced. If I can search for “reviews San Francisco chowhound azie” — and I can — and get an aggregate thread of web-wide posts and information, including stuff that doesnt happen to be posted on Chowhound’s board, why would I want to wade through this? I don’t.

I think I finally understand why “blogging” _might_ be such a big deal.

Check back shortly for “wink answer mov video converter“.

I just ran across agkamai’s Cat2Tag plugin for Wordpress — looks great. My first suggestion would be to add type-ahead tag guessing like on the del.icio.us post interface.

I’ve been thinking about tagging UI as well, and this discussion about tags, commas, spaces, etc got me thinking a bit more today. For one, I tend to agree that spaces are bad delimiters, and I agree that “what’s wrong with commas, anyway?”. More importantly, however, is how to get users to understand how to use a delimiter. As others have noted, they do understand the comma as a delimiter for email addresses, but search engines have trained us to use spaces to seperate words in free text fields.

So, here’s an idea - what about a dynamic UI that provides some visual feedback as I type. If I type san francisco without any delimiters, each word is highlighted seperately (two “tags”). A visual cue (perhaps a comma fading in to the right of the text field?) reminds me of the delimiter. If the user then adds a comma — san francisco, — then the highlighting changes to show that the words are treated as a phrase, or single “tag”. Here’s a crude illustration of how I’m thinking it would look:

I’m sure someone can figure out how to do this sort of word-by-word highlighting inside an HTML form field…

I was quoted in the Wall Street Journal today. Of course, my name isn’t exactly Bowen Delle, but that’s pretty close — and not nearly as bad as some other corruptions I’ve heard.

So, why bother with “tagging”, or a service like del.icio.us? Who, you might ask, has the time to think up and key in a bunch of descriptive keywords for every bookmark. Just bookmark it and… forget it!

Exactly.

I look in on my old IE or more recent Firefox bookmarks very rarely. In fact, there’s only one Firefox bookmark that I use with any regularity (aside from those in my “Bookmarks Toolbar”) — and that’s for the status page of our WiFi router. I don’t use the others because I’ve simply forgotten about them - they’re buried deep in folders, locked on my hard drive, invisible.

Joshua makes a nice point in his interview with Jeremy Wagstaff: “Search is more associated with the recall, whereas tagging is more associated with the storage.” Quite right, because of course an individual doesn’t really have much opportunity to “store” things in general on the internet - after all, your blog (if you have one) is just one of several billion web pages. Search is broad yet shallow, in terms of personal differentiation. Bookmarks, whether on your hard drive on with a service like del.icio.us, are narrow but much deeper, richer, more satisfying on a personal basis. You can’t find anything and everything on the internet in your bookmarks, but you can (hopefully) find those things that you care about most.

So, why use del.icio.us (or one of the others) instead of your browser’s own bookmarks?

1) Central storage: A online bookmarking service is accessible from anywhere you use the internet, not just one PC. This applies to any of the services, including toolbar-based bookmark services like Yahoo and A9. Del.icio.us can be intergrated into Firefox with Foxylicious and various bookmarklets. Someone could also easily build a del.icio.us toolbar.

2) Personal and global view: An online service lets me see not only my own bookmarks, but those of other people. In data-architecture terms, I can pivot on any one of at least three axes: person, URL, and tag. Just for starters, I can see who else has bookmarked the URL that I did, what other URL’s that person has bookmarked, and what other bookmarks have been tagged with the same tag. This is social networking with some real utility. Friendster is fun and all, but once you’ve gotten all your friends to sign up, there’s not much reason to go back unless you’re looking for dates.

3) Integration: Like many other new-school web services, del.icio.us has open API’s, which allows people to build services that talk to and leverage del.icio.us. Contrary to the prehistoric thinking of some developers, your worth is not measured in how much you keep others from knowing — it’s measured in often what you know is used. Check out some of the experiments floating around del.icio.us.

4) Tagging: Although I came upon the idea by way of del.icio.us, tagging can be applied to almost any application, including browser-based bookmarks. The advantage with an online service is that I can pivot on tags not only within my own data, but across other people’s bookmarks and even across other services.

More on this later…

PS - AAPL is up to 73, which means I’m almost to my Mini. Now lookee there — I wonder if Apple chose the name in part for the affinity with the other Mini. In the meantime, I’m reading up on Mac software… Also, they’re not exactly rare any more, but I have plenty of Gmail invites - let me know if you want one.

Jeremy Wagstaff of WSJ/Dow Jones called today, asking what this “tagging” thing is all about. Remember the 80’s? Tagging was something else entirely then. Par for the course for a new-school journalist, he interviewed me via Trillian from Indonesia, and asked if it was OK to blog about our conversation. That prompted me to revisit my blog, since I’ve been somewhat OTA for a while.

So when I went and logged into WordPress, I found a post that I had drafted a few months ago. I had already been writing about tagging and tag-centric views, and having just been down at the Technorati Hackathon I was musing about how trackback is an imperfect pivot for following conversations:

I posted last night about the hackathon, and linked to the wiki. Technorati shows only one other person linking to that URL — so where are all the hackers? Is looks like more people linked to Dave Sifry’s announcement page…. So, which URL to include in my post, so that it’s properly connected to the Technorati Cosmos?

This is of course quite relevant to the topic of the day. Trackback sucks because even though URLs are good unique identifiers, they don’t carry enough semantic weight to serve as useful pivots in information space. Technorati is a partial solution to the trackback problem, because even though it provides a central place (an anchor for pivoting on URL’s), if I’m writing about a topic and I want to link to other conversations on that topic, linking to one and only one URL that might have to do with that topic is both insufficient and cumbersome. By linking to the URL for the Technorati Hackathon Wiki or to Dave Sifry’s announcement of that event, I was asserting that my post was related to one or the other (or both) of those specific entities, thereby excluding reference to anything else.

On the other hand, if I include (as I did) a reference to a broader concept, such as my “technorati” tag, then I can easily pivot on that tag, linking to other entries [throughout the web] on that basis. That’s what I was writing about back in April of last year, that’s why I started calling my WordPress categories “tags”, and that’s why I pulled together blog entries and del.icio.us posts, and then eventually photo tags into a tag-centric navigation system on my blog. I’m sure I’m not the only one to have come up with this — but it does look a hell of a lot like Technorati’s tag aggregator.

So, what’s with all the hype? Tagging isn’t anything new — it’s just simple meta-data, applied in a flat (vs. hierarchical) namespace. It’s basically open source for information classification; the opposite of both the brute-force computational approach (”AI”), and the top-down, fewer-experts-know-best method (libraries, Yahoo in 1996). I’ve always thought that rigorous taxonomies are not “profitable” in terms of information management; they cost more to maintain than the value they add minus the error rate. It might be true that taxonomies can only be maintained successfully as public goods, e.g. the Dewey Decimal system, the domain name system, etc. Google leapt ahead by combining the machine approach with some of the benefits of the ‘Wisdom of Crowds’: PageRank. But this is still implicit/implied vs. declared data. That’s what it comes down to: can you reach a more accurate understanding of what something is about via interpretation of the content itself, or by examining the aggregate of what author(s), readers, and commentators declare?

I can also understand the point of view of those who think that it’s optimistic and naive to believe that aggregate amateur metadata is of real worth. I haven’t had to time to digest all of the discussion on the topic, and there’s probably something I’m missing in their criticism. But, my own experience — not just in blogging, but in application and data architecture in general, tells me that there is real value here. delicious is certainly a testament to that.

So it’s not technology — it’s the applications. For me, it was a combination of blogging and delicious that brought the utility of tags to the fore. Once I saw applications like Flickr and gmail using the same concept I began to think more about how to bulid further. I have a few more ideas on the topic; it should be interesting to see what happens over the next few weeks. I’ll try to keep more up to date, as “tagging” is clearly exploding.

In the meantime, let me be one of the first to cast a vote against the horrible, awkward term: folksonomy. Really, folks, we can come up with something better. We certainly should.