Web 2.0 needs Data 2.0, Data 2.0 needs MetaData 2.0

November 26th, 2005, By Duncan Gough

If it looks like I’m obessesd with the idea that Web 2.0 needs Data 2.0 then it’s forgivable, given that Millionsofgames.com is the only casual games Web 2.0 application. It’s also the only website that allows users to bookmark and tag games.

As They Might Be Giants put it, not to put too fine a point on it, but Millionsofgames is very, very unique. What sets us aside from other tagging websites, like del.icio.us and digg.com is the way we aggreagate user data and use that to form a ranking of each users’ worth to MOG, giving us a score table of user rankings.

With our background in games websites like Playaholics and Chickstop, making scoreboards out of user activity is second nature. Chickstop has a Top Chicks list and Playaholics has an Ultimate Playaholics one. However, given the wealth of data available to us on Millionsofgames.com, the user rankings are much more complicated. Making a game out of user participation, though, was a simple choice for us. Someone noted recently that Flickr grew out of an online game, a browser based MMO called GameNeverEnding. Sadly, the GameNeverEnding did, but we got Flickr instead. Over on k5, the idea of turning user participation into an MMORPG was recently floated. Again, it’s the amount of data that these types of user-generated websites that is so valuable. And it’s this wealth of data that I really wanted to talk about.

It’s a simple point really – there’s is so much data supplied by users when they upload a photograph to Flickr, tag a website or click a link to visit that bookmark on del.icio.us, play or rate a game on Millionsofgames. Inevitably, only a fraction of that userful data is being used at the moment. The obvious stuff – the boorkmark url, the user comment, the tags to generally categorise the user-supplied data. However, as I’ve watched the user data pour into Millionsofgames, I’ve begun to realise how much of it is being lost. It’s as if the white noise that I thought was the sound of lots of lovely web traffic has turned out to be the sound of lashing of user generated content spraying out of the leaky pipe that is a Web 2.0 application.

For example, there are two ways to MOG a game on Millionsofgames. One is to MOG a game that is not in our database. Doing so earns you ‘followers’ once other people start to MOG it. The other way is to do just that, MOG a game that is already in our database (thereby becoming a follower of the main MOGGER).

The same is true for almost all social bookmarking websites. In both cases, there is a key piece of metadata that none of them are collection. Namely, how did I get here?

If I’m bookmarking Jewel Thief for the first time on MOG, it would be a huge bonus for the website to know how you got there. Did you Google for a ‘free online game’, find Playaholics and then navigated to a game in the Top 10 games list? How about a game that I MOGGED from within MOG? Say I was searching for games on MOG, browsed the POP list of games, browsed that list according to the ‘Arcade’ tag/keyword and then found Jewel Thief in there? Isn’t that meta data hugely valuable?

Given that social bookmarking websites are about making it easier for the user to find websites they found userful previously – given that the social bookmarking website offers tags and a search engine to let you find things quicker than if you had to search you browser history, isn’t this kind of meta data – how did I get here – incredibly valuable? If I lose my house keys and try to find them again, I retrace my steps (or snap at the first person to say ‘where did you last see them?).

Finding bookmarks is no different, if I played a great game the other week, I should we able to start searching my bookmarks for games MOGGED last week. If I played a game last week that I found via an initial Google search, I should be able to search my MOG list using that criteria. Just as I should if I played and MOGGED a game last week that was already MOGGED. All this metadata is valuable. however not all this metadata is accessible and that which is, I’m not sure how to present it to the user.

As ever, del.icio.us has started down this path by subjectively guessing what tags you’d like to use when you bookmark a url that is already in their database. It’s very clever already, but we need to add more metadata to tags, like how you arrived at this item – if I tag an item already in del.icio.us or millionsofgames.com, then it would make sense to record the path I took to find it. It could be a search query, or an act of navigating around the website. Since there are so many views available on tagging sites like delicious and millionsofgames, knowing that I found a game via suttree’s most popular games list rather than the alphabetical list of games has some significant value.

Just as, with the onset of tabs in website design some time ago, there was an initial delay before developers began to save state so that each tab was relevant to the page you were viewing (e.g., the scoretables on Playaholics attempt to show each tabbed time period relevant to the game scores you are already viewing), we need to maintain and save the state once you commit to bookmarking a game or a website. Back then, every tab was dumb and didn’t really take notice of your path throughout the site, making tabbed website designs quite frustrating. At UpMyStreet we introduced them and made sure they maintained state at the same time as a number of other sites, such as the BBC, followed suit.

Of course, how I do that and how I make the representaion of that data useful, understandable and worthwhile, is the next problem. For millionsofgames.com, I’m tempted to quickly hack something in using the screenshots – maybe just a screengrab of the breadcrumb trail and the url (since they’re RESTful). That and microformats, of course.

2 comments

« One interview, one reviewCasually Microsoft »