Thursday, June 30, 2005

Open Letter: major UI issue

I love I wish Joshua Schachter all the best, and as I understand it, "the best" means everyone using his wonderful creation.

Joshua, can you please, please fix one UI problem that is (I suspect) costing you thousands of users and hundreds of thousands of uses every day?

How life should work:

When I go to Google, and I want to find something, I find a box that has a blinking cursor in it, and a grey button titled "Google Search". Unless I am waaaaaay behind the curve, it dawns on me (me being the zero-sigma Internet user) that perhaps I should type what I want, and click "Google Search" or even (gasp!) press (Enter).

How works:

When I go to, what do I see?
1) A big list of tagged items that I may care nothing about
2) A list on the right side of "most popular" that I probably care at least slightly about
3) A few mysterious text links in the top-right: "bookmarks index post settings about"

How do I find something?

I dunno.

Oh, it turns out, I go up to the URL entry field on my *browser* and type /tag/something to get what I want. And if I want "interesting blog" then I tyle /tag/interesting+blog.

Gee, that was obvious!

What if I want to see someone's bookmarks? Again I go to the browser's URL entry field, and add /onohoku/ or /fred/ or whatever (there seems to be no fred or john on but of course there's an /ethan/...miller... we Ethans are early adopters!

Now this being Web 2.0, of course I could hack together my own bad-ass interface for del.icious and release it to the world. But that would not solve the core problem, which is that right now, some frustrated newbie is going to the main site, laden with glorious potential, and clicking away frustrated because he damn well can't figure out how to *get what he wants*.

Can we please, please, have a text entry box or two? Or even (gasp) a single line of text explanation of how to get what we want?

It will make the tagging universe grow *ever* so much faster...


Wednesday, June 29, 2005

Google Maps API is launched

Read about it here. In my last post I talked about Google and Yahoo trading punches - well, I think that Google is landing fewer but bigger shots. Google Maps is a roundhouse, and until Microsoft launches their TerraServer product as Virtual Earth on MSN (supposedly launching any moment at Where 2.0) Google Maps is the standard. There are already all sorts of cool mashups out there, and now that the API is out, this will explode. Um, where is Yahoo on this?

Flickr takes Yahoo Social: My Web 2.0

Well, along with everyone else on the small planet known as Silicon Valley, I've been wondering what Stewart, Caterina, and the Flickr folks have been up to since Yahoo bought them... since they certainly haven't been updating the Flickr interface ;-)

Today Tyler drops me an email saying "Check out Yahoo's new My Web 2.0, it's social search and it's cool." And lo and behold, who's writing the Y! Searchblog section for MyWeb but Caterina...

Damn, it is cool. It's delicious Yahoo search.

So what does this mean? Well, there are four fundamental ways to ways to determine relevance in search results. These are:

1) Market/collective/algorithmic - Call it the law of large numbers, the wisdom of crowds, or statistical truth, but PageRank and similar "hubs and authorities" algorithms figure out, on average, what everybody thinks of a given site.

2) Expert/editorial - Roger Ebert on movies. Julia Child on cooking. Henry Kissinger on diplomacy. They've spent a lifetime figuring out what is the best, and the'll tell you so. Yahoo's roots are in this sort of directory/taxonomy/recommendation based organization.

3) Personal - Based on the user's past history and demonstrated interests, give them results that most closely match their tendencies. Amazon's A9 is the best current example.

4) Social - Based not only on the user's past history and demonstrated interests, but on those of all the people they know. This is a fascinating middle ground (possible sweet spot?) in between personal and collective/algorithmic -- you've got enough critical mass of opinions on key topics to hopefully eliminate outliers, but you've got a close enough connection to the user that you can bend your results to match their point of view.

A note on tagging: Tagging doesn't change any of the above four categories. The voting mechanism -- bookmarking, visiting, linking, tagging, reviewing, whatever -- which is being rolled up in any of the above methods is entirely (OK, largely) independent of the method used. Flickr is all about tagging, but don't forget that they started out as a social network -- and that social sensibility is what they've brought to My Web 2.0.

Google and Yahoo are introducing products like two boxers standing toe to toe and punching. Google is throwing stunning roundhouse technologies like Foreman, shaking everyone's perception of what's possible whenever they land a punch. Yahoo is jabbing and dodging like Ali, clever sneaky punches that do a lot of damage but don't cause the crowd to Oooh in quite the same way.

Ali beat Foreman the second time they fought.

Update: John Markoff in the NYTimes is all over this. He gets the fact that we're talking about the re-creation of local media, too. The curtains just lifted on the future, folks, and the play is about to begin...

Update Update: Battelle, as usual, is all over this first. If you want any more proof that all of media is changing, the fact that his book *still* isn't published (all of, what, three months later?) drives him to asterisked profanity. And to answer John's implicit question, yes, communities of experts will emerge who will tag and organize for a living. They'll be the modern-day equivalents of Ebert, Child, and Kissinger - but they'll express their expertise not in books, columns, and Wall Street Journal op-eds, but tags, blogs, social networks, and other forms of relevance votes, connections, and organizations of information. tb

Tuesday, June 28, 2005

"Inbound" vs. "Outbound" definitions: Vertical search is horizontal!

One of the questions today at the SDForum Vertical LEAP conference has been "what is vertical search?

There are two simple answers to this. Firstly, there's what I'd call "Inbound" vertical search. This means that anyone in the world, at some given point in time, wants to do a certain thing:
Get a job

So the "Inbound" is people (temporarily) participating in a certain type of activity.


Second, there's the "Outbound" concept of vertical. This is the sort of vertical that business-to-business publishers like Primedia and CMP and Reed Elsevier have been working on forever. Oh, and there was this cool "bricks and clicks" company called eBusinessMedia...

Here, the vertical is based on "who I am" rather than "what I do". The music business is a vertical. Agriculture is a vertical. Medicine, with several sub-verticals, is a vertical. Music, agriculture, and medicine all need employment/jobs services -- all need secondary markets for used capital equipment -- all need dedicated industry news sources -- but the "vertical" here is the fact that medical jobs, markets, and news are entirely different from agriculture jobs, markets, and news.


Oddly enough, from the standpoint of this traditional view of "vertical," the distinctions that are driving this conference are actually horizontal.


Hmmm. Leave it a bunch of technologists in Silicon Valley to mix up their X and Y axes ;-)

Local Media, Technorati, and Vertical Search

I am sitting here at SDForum's "Vertical LEAP" vertical search conference, where I've sat through a local search presentation from Google, Yahoo, Ask Jeeves, and MSN which pretty much missed the point entirely. I asked a question about a) events and b) local search as local media, since neither (enormous, gigantic, huge) topic had been mentioned by any of the assembled gigantic players. Credit to Brady Forrest, PM at MSN for their local search, for at least mentioning mobile -- yet another (enormous, gigantic, huge) topic they'd entirely slid past for a whole hour. If these guys get it, then they're holding their cards very close to their chests.

I'm working on an events search company called Zvents, which heretofore is unannounced. So there. I announced it. Secret closed demos on the web in a month, live beta by September, or faster if Tyler starts adding Red Bull to his coffee.

So I am spending a lot of time thinking about local media. And none of these local guys get it at all, and at the same time, Technorati (hats off to Dave Sifry!) gets it in a (enormous, gigantic, huge) way. I'll be the Nth guy to link to the Live8 page, and the second or third guy to link to Peter Caputa's discussion of that page (right on Peter, and we should talk about events) and I'll even throw in a bonus link to Chew Shop on Caputa on Technorati.

So what do I think?

Well here are a few relevant facts.
1) Local media is enormously fragmented. There are 70 newspapers in the Bay Area, plus tens of TV and radio stations, and they are *all* supported by ad dollars.

2) Classified ads were a nice chunk taken out of local media, but those dollars are largely gone; high-value categories like jobs, housing, and for-sale went to Monster, eBay, and the like long ago.

3) The three legs of local media are 1) reporting 2) editorial and 3) advertising. Let's take these in order.

Reporting: When we look at a project like Dan Gillmor's open media / bayosphere / We the People, you see the foundations of local reporting. Remember that this is not just going to be people doing this in their spare time -- we are going to get latter-day Ben Franklins running around researching stuff and writing it up (or podcasting or vblogging it) and publishing it online. Reportage = conceptually sorted, 20 years of implementation to go.

Editorial: I have a whole separate riff about how political blogging and tech blogging represent the explosion of the vertically integrated media model -- guys like Kevin Drum or Joshua Marshall provide me with more editorial insights than does the New York Times; but for the moment, they are reliant on the "MSM" mainstream media to provide the basic grist for their editorial mills. Most interesting to me is the extension of editorial blog voices beyond the narrow reaches of politics and emerging technology. I firmly believe that the reason these "went first" is that the foundation of bloggable news was firmly in place in textual form, and the blogging tools that sprang up made manipulation of that text very simple for bloggers. Myspace is actually an example of music blogging, I think -- 18 million people go there because, fundamentally, they gave indie bands the (different) tools they needed to post MP3s, talk about their shows, and post their band/brand image online. I think that with the emergence of "structured blogging" standards for other categories -- shopping, events, finance, and so forth -- we will see and explosion of new blogging editorial categories as, for instance, a financial blogger can quickly and easily point to financial data and charting information to make his (her) points.

Advertising: This is the part of local that has gone the furthest, and the least; local classified (in certain categories) went 10 years ago, but local newspaper display, local TV and radio, and lots of classified -- plus flyers, posters, billboards, etc. -- has yet to go. And we are talking tens of billions of dollars. It will follow the eyeballs, which will follow the tools, editorial voices, and reportage that is already moving to new media.

Technorati gets the media thing, as demonstrated by the Live8 page -- but it's not clear that they get the local media thing, which fundamentally has shaken out to be an oligopoloy the first time around between Knight-Ridder, Hearst, Tribune, Gannet, etc. with very little local competition. Will it happen that way again? I'm not sure - and this Technorati thing may be the emergence of a new NBC or a new USA Today, not a new Knight-Ridder -- but what I'm sure of is that none of the local papers get this, none of the big portals get this, and there is money laying all over the table to be scooped up.

Game on.

P.S. How much has the world changed? This places is full of VCs and entrepreneurs and the nicest thing that's happened to me today is that Marc Canter linked to me in his blog. Thanks, Marc!

Monday, June 27, 2005

Response to Kevin Drum - Three Thoughts on Future Media

In light of the recent Supreme Court decision on file-sharing, political blogger Kevin Drum asks:

"The year is 2015 and Columbia has just released Spiderman 7. The next day, 10 million people with no technical savvy at all go to their computers, stick a Blu-ray disc into their DVD drive, log on to Movies4Free... Three minutes later they have a 100% perfect DVD...As bandwidth increases, DVD technology improves, and software becomes as easy to use as a toaster, every piece of digital content on the planet will be available within minutes...Is this OK? Or do you have a different vision for the future? If profit-based movie/music distribution becomes essentially impossible, do you think the content industry will somehow adapt and get its revenue elsewhere? Or will content creators continue creating but just make a lot less money at it?"

Three thoughts:

1) Right now, radio broadcasters make more money from music than music labels do. U.S. CD sales are $17 billion, radio advertising is $20 billion. Broadcasters' ability to take the labels to the cleaners for the past, oh, ninety years is due to a court decision in the early days of the 20th century that set the recompense rate exceedingly low. The NAB has had enough political clout to keep it there ever since. Cool, huh? Note that they make all this moolah by... giving songs away for free.

2) Before Caruso (correct me if I'm wrong on this) there were no recorded-media stars, and instead there were many more local peformers of moderate skill who played directly to the people (y'know, live) and made a decent living doing so. If the hyper-promotion of mediocrity machine that is the labels goes away, perhaps we can move somewhat back in the direction of this former model. Quality will still rise to the top; with free distribution of songs, it would do so as a matter of course. But if the primary payment mechanism for musicians was for *playing music* instead of recorded album sales, then, it would be terrible! No one would get more rich than Jerry Garcia. Have you ever seen Jerry Garcia's house? That sounds OK to me.

3) Movies are a very social medium with extremely high costs of production. Right now, they emphasize cost-of-production over, say, quality. I get better acting at my local theatre than I do in watching a Jerry Bruckheimer $200m extravaganza -- but the "production values" of the local theatre are a bit lower. If production technologies get cheaper (and they will!) then the local theatre can start to compete -- meaning that we'll get a much wider base of content producers, similar to today's documentary situation. Unlike music, it's not unreasonable to hyper-encrypt movies during their initial "theatrical" phase and only show them there -- people seem to like watching movies together, and after you've made back your (much smaller) costs of production plus some profit, you throw the thing into the maw of the collective and start working on your next piece. Again, surprisingingly similar to the local-theater model, which seems to work just fine.


Tuesday, June 21, 2005

So what about tag spam?

Tyler asks me another question:
"I'm a little concerned about tag spam. Is this a problem on Flickr and other sites that allow communal tagging?"

So I went and did a search.

I haven't found any discussion of it as an issue. Comment spam on blogs and comment trolls, on the other hand are well-known and much-discussed problems; none of the solutions are perfect, but include:

- Registration (know who posts something)
- Moderation (increase or decrease visibility by editorial intervention)
- Deletion (the ultimate in editorial intervention)
- Complication (making spam hard to do quickly and automatically)

Tag spam might be different than blog comments, because:
a) You can't include a URL (which spam needs to do in most cases to be effective)
b) You can't express yourself to others in a personalized way (which most trolls seem to want to do).

So it may be anodyne enough to not really be useful to vandal communities. Just a guess.

Some related links:
A loooong but fairly interesting post about tags. No mention of spam. John Dvorak pans tagging in PC Magazine. Some brief thoughts on vandalism in reply to Dvorak -- apparently has had some spam Another reply to Dvorak. A flat-out awesome post on the meaning of RSS

Are Tag Clouds Useful? And how to make them more so

Tyler asked:
"Is there a better way to display tag utilization than the tag cloud?"

I would say yes. The "tag cloud" was a cool concept but I find it of limited usefulness. Its axes of meaning include:
1) inclusion (vs. exclusion - the cloud is finite)
2) font size
3) order

I have seen clouds ranked both alphabetically (another one) and ranked by use; when listed alphabetically that means that there is only one axis of meaning (aside from the set limitation of inclusion) which is font size.

It would be possible to include font color as another axis, and even bold vs. italic, but what does that tell the user? It will not be clear.

Absolute popularity is a moderately interesting feature of tagclouds that, frankly, is best represented by a list; I assume that the reason that tagclouds were first done is that the resultant visual mass fits more neatly on to a web page than a 100-item vertical or horizontal list.

Here is an idea that might make tag clouds more useful: Create an actual honest-to-god 2D map representation of them. Sort of like a network diagram, but all node and no connection between node. User types in a tag -- "chocolate" and you create a custom 2D map based on "events that include tag chocolate also include these tags" and then let the user scroll around on that map or click on that map, in Google Maps fashion (can you tell that I have been profoundly influenced by Google Maps? ;-) and get to "food" "wine" "laborador" "ghirardelli" "roses" or whatever, and gradually scroll away to the outer bounds of the related-tags-surface.

Now what you have is a tagcloud that shows not overall popularity (a "who cares" feature -- when was the last time you checked out Google Zeitgeist, much less derived some useful information from it?) but relatedness to something the user cares about, presented in a navigable visual form.

Might be interesting.