Tag me
: Yahoo puts up its new save-this/tag-this beta.
: David Weinberger points to a service that lets you tag books. Well, of course: What happens when we don’t just tag pages but things — movies, destinations, books, gadgets… And what happens when you can get to that on the road (in the bookstore, you look for recommendations of books about…).
Of course, you can tag people.
: I have been showing everyone who’ll listen Flickr’s “interestingness,” a secret-sauce algorithim that, I believe, uses vectors of interest and link and social patterns to find the photos Flickrers find interesting. And they are interesting.
What’s great about this is that it exposes not the wisdom of the crowd but the taste of the crowd.
October 28th, 2005 at 10:22 am
President Al Gore’s Exiled Speaker of the House — Sean Penn!
Sean Penn’s Blog
Al Gore’s Blog
October 28th, 2005 at 10:51 am
Hello! I’m the creator of Reader².
I will be glad to hear any feedback that can help me to improve the site for better usefulness and usability
October 28th, 2005 at 11:07 am
Coming Attractions: Net Porn, Brown Power.
This blogging thing is thirsty work. Give me enough Premoistened Lemon Pledge Wipes, and I’d rather clean the Augean stables. I’ve been under the hammer of a dozen deadlines, hence my absence from the bully pulpit. That’s my cover story,…
October 28th, 2005 at 11:40 am
Jeff, the link to ‘people’ is wrong.
October 28th, 2005 at 11:57 am
Hi Jeff;
It seems to me that tagging will eventually (and in short order) be applied to *everything* because it gives language a globally useful new application.
Language is simply a way of tagging our common experience with words. But online, tagging means every word can become an information stream.
Language 2.0 perhaps? (Sorry, couldn’t resist. It seems everything sounds more hip with a “2.0″ tag
I was personally awestruck by how tagging gave rise to spontaneous emergent democratized real time reporting during the London bombings.
But as tagging converges with our environment through wireless access and applications like Google maps, tagging will create sticky location-based fountains of useful knowledge. It’s possible now through GPS and it’s growing family of location-spotting cousins to tag areas of about a 5 foot sphere horizontally and vertically. That means the very table you sat at for breakfast or the hotel bed you fell asleep in. Probably scenic views from your plane seat. Or anything.
Keep up the great work Jeff. Don’t know how you do it.
~ted
October 28th, 2005 at 12:55 pm
More than just “interesting,” “interestingness” will be the way that Yahoo! reclaims a little piece of search from Google. Today image search at both Google and Yahoo! is largely broken. Do a search for “San Francisco” at both Google and Yahoo! Image Search and you will find a hodge podge of mostly mediocre images.
Now do an image search for “San Francisco” and rank it by interestingness on Flickr and you get a substantially superior line up of photos.
It will only be a matter of time before Yahoo! integrates Flickr’s algorhithm into their image search and Yahoo! will begin serving up much, much better image search pages than Google, this could potentially cause a buzz and a shift in image search traffic from Google to Yahoo!
The main advantage Flickr has with interestingness is that it is based on human input as opposed to pure computer generated search algorhithm. The actual human ranking, rating, tagging, commenting, etc. that helps determine interestingness for Yahoo! will give them a leg up in determining superior images.
On the other hand, working against Yahoo! is 1. They have been slow to integrate Flickr’s algorithm into their image search (I have no idea why, it’s such a no brainer) and 2. Google image search results are actually improving. Although I doubt Google still has any “human” element to their image search page rank, I do believe that they may be targeting certain photobloggers for instance who consistently produce quality work and enhancing their image rank. Not that my photography is great, but it is probably better than average from an image search result and most recently with the Google reindex a few weeks ago my Google Image Search traffic has trippled. I’ve also noticed in doing Google Image Searchs that since the reindex three weeks ago that their images are in fact better than pre reindex. They still are not as good as Flickr’s interestingness ranked searchs, but they are better than they were before.
October 28th, 2005 at 12:59 pm
I’m working on the idea of centralization and interestingness. It’s easy to create a page like that when one big Flickr database contains all of the information. How do you make that work in a decentralized system where everybody puts their photos online in a P2P like environment? It is possible and it means we’ll get the ad revenue, possibly by creating a new type of file format that stamps an ad on an image file and includes a link to the advertiser.
October 28th, 2005 at 1:13 pm
Kirk, having the user generated data that creates interestingness for flickr is priceless. It gives them a huge leg up. I think it has to be done in a centralized place — but maybe not. The key is to somehow collect the human responses to photographs (views, favs, comments, activity, etc.).
Alternatively you could possibly create some kind of a system that ranked photobloggers and photographers instead of their individual photos and then automatically assign higher rank to the better photographers.
Photoblogs.org, Brandon Stone’s site, for instance has a mechanism to vote for photobloggers. Theoretically I’d suppose, someone could mine the top rated photoblogs for search terms and increase their image search page rank thus producing superior results. I’m not sure how/if this could be done on a P2P basis, although I’d suspect that certain photographers could develop reputations that could influence rank. Of course you’d need to get photographers to play ball with their work and images in a P2P environment (copyrights, etc.). Flickr at present has very favorable terms towards it’s photographers owning their rights, creative commons licensing, etc. that seem to work.
Trying to do this for ad revenue though may be increasingly difficult. Already even Flickr is coming under scrutiny (see this post: http://thomashawk.com/2005/10/flickr-caterina-fake-anil-dash-wealth.html for more on recent criticism leveled at Flickr for ads from Anil Dash).
October 28th, 2005 at 1:46 pm
You know, you could take, say Ares Galaxy, an open source p2p app and modify it to have a Flickr like user interface appearance, add metadata to the image file format and have a free, open source alternative to Flickr. It might have a catchy name like OINk for Open Image NetworK. It could be the second sourceforge project I start that I don’t have enough time to devote to
Maybe I’ll make time.
October 28th, 2005 at 1:59 pm
Kirk: I think the tag data can work in a distributed fashion with either (a) data sharing or (b) microformats that tag the tag data. So…. You can go to Technorati and get a feed of tag data. You could probably get a sense of vectors of tag usage. But I’m not sure how you’d get the social relationships that also matter.
October 28th, 2005 at 2:07 pm
The social relationships are what you need though in order to determine rank. The interestingness algorithm is dependent upon the central database of collected info on views, favs, comments, tags, etc. It is indispensible to providing the superior search. Unless you are going to pay a bunch of college students to rate and rank by the hour which would be too expensive and time consuming you will need the social relationships.
My only thought would be to rank P2P users themselves somehow and to then automatically give their image rank for file titles and tags some kind of a multiplier effect. I.E. If user X has consistently downloaded files then they are probably more likely to be consistently good, thus when they mark a photo as “flower” it should be given priority over others’ flowers. Even if you could build this though you have to ask yourself, will they come. Much of why Flickr works is due to the vast number of ways for the users in the flickr community to relate to each other. It’s responsible for the adictiveness of the site which is what prompts participation.
October 29th, 2005 at 12:09 am
Wow, interesting discussion… I read the comments, brainstormed, and came up with a way to create a system that would rival Flickr and Digg with a few hundred lines of code and no cash, which happens to solve the ownership issue (I think).
But first, here is a feature comparison from both sites:
Flickr
- Content - user submitted (images)
- Tagging - proprietary internal tagging system
- Discussion yes (per image)
- Relevance filter - interestingness based number clicking the star(favorite)
http://flickr.com/photos/tags/dog/interesting/
Digg.com
- Content is user submitted (links) -
- Tagging - no but it does have categories and a full text search of the link description
- Discussion (per link)
- Relevance filter - people clicking “Digg This” moves up on the page
Del.icio.us
- Content is user submitted (links)
- Tagging YES
- Discussion (per link - show only submissions with tags AND description, newest last)
- Relevance filter - number of people tagging - popular page and per tag popularity - http://del.icio.us/popular/dog
So here’s how you could make Flickr irrelevant for $0 (not including domain name)
Flickio.us?.com
- User submitted content (links to photos hosted anywhere including Flickr)
- Tagging - uses Del.icio.us API similar to how Flock prompts you for your un/pw so you’re using Del.icio.us but might not even know it - less friction.
- Discussion (per image link) extracts and the del.icio.us description feature.
Relevance filter - uses Del.icio.us popularity instead of interestingness but with the same effect.
You could even link to Flickr images and simply not use their interestingness or discussion features to build a critical mass. There is nothing magical about the Flickr Interestingness system even though they’d have you believe otherwise. It’s just a lot of brains making decisions combined with a critical mass of images.
The semantic web would be useful here, if someone uploads to flickr and don’t make something private the assumption is you can repost the image elsewhere. That’s where the Semantic Web comes into play. image metadata needs to include things like the license, watermark, description, some static tags, fee for purchase, etc.
Some predictions:
- Flickr is about to redesign their homepage so it looks like Digg.com. Digg might get into the photo business without having to host the images, it could create a demand for better open source image hosting services - Flickr would prevent hotlinking no doubt.
- All of these user driven web 2.0 apps are going to start using a common source for tagging (Del.icio.us API) and maybe even discussions because they can borrow the critical mass of users (think trackbacks in blog comments but that are aware of the previous trackbacks). At that point people are going to say, just like they are now with Flickr, why is Del.icio.us making money from my tagging effort? Which leads me to the third prediction.
- A Del.icio.us backlash leading to calls for a non-profit tagging system will linger in the back of the mind of Jimmy Wales of Wikipedia fame. The WikiTag is unveiled and people cease to make money from the wisdom of the crowd except for the individuals in the crowd (advertising metadata - abusers would be subject to the wrath of 1 billion “spam” tags when the blogoshphere got wind of copyright/creative commons license violations). Try typing in Spam in Del.icio.us and rank by popular, now imagine that evil-rank embedded in a pagerankesque feature built into every web browser. The google pagerank doesn’t understand infamy, it just knows popular - take the reddit system - stir in some tagging and you’ve got a Google killing recipe, it’s already happening to their image search.
- This is all going to happen within a year.
October 29th, 2005 at 2:22 am
I think tagging will ultimately lead to a new form of networked & engaged community of users and content. Ultimately am expecting tagging to help in connecting a network of interested prospects. Tags will become the pipe connecting users of similar interests. When google starts selling its print and TV ad model, a new engagement model will get created. That’s where I see tags making a huge impact. Read more of this in my blog on google and engagement model.
October 29th, 2005 at 5:18 pm
In response to your question “What happens when we don’t just tag pages but things…” that’s why we created Clipmarks.
October 31st, 2005 at 8:17 am
[...] But second, in an interesting comment discussion under my post yesterday on interestingness, KirkH asks whether interestingness requires that the content judged be hosted on one site. That’s a good question, for interestingness appears to be about both vectors of interest and also about relationships and I’m not sure whether or how the data to feed that algorithim can be done across a distributed network. [...]
June 15th, 2006 at 5:43 pm
[...] Says BuzzMachine: "What’s great about this is that it exposes not the wisdom of the crowd but the taste of the crowd" [...]