Sunday, January 01, 2006

Tinfinger screenshot 2

At the right there you can see the second of the screenshots from the forthcoming beta of Tinfinger, the human search engine. The profile pages are where the action is on Tinfinger, and this is only a taste of what will come.

The person profiled on this page is an Australian Rules footballer by the name of Nick Riewoldt. Everything on the page is live from our database except the blog entries on the right (not many people are blogging about Nick at the moment). The drop-down menu at the top has human, category and tag as options. There will be further functions in the right column when the page comes out of beta. Google AdSense's poor old spiders don't know much about our pages as they're hosted on our test server, so they've looked at the title text and honed in on the "tin" part, evidently!

I guess if you were reminded of Memeorandum by our first screenie, then this one will remind you of Wikipedia. That's not a bad thing. Where Tinfinger hopes to expand the Memeorandum concept beyond tech and politics, we hope it will effectively become a "category killer" for Wikipedia by focussing particularly on humans. Wikipedia is to the Encyclopedia Britannica as Tinfinger will be to Who's Who, as I have said elsewhere. Or at least, that's the plan. And like the EB was, Who's Who seems to me to be over-ripe for updating to the 20th century, let alone Web 2.0. Who's Who Online looks like it was designed in 1995 and hasn't changed a bit since then. Marquis Who's Who and the International Who's Who are better-looking but their content is still behind a pay wall. The International WHO'S WHO Historical Society is even worse - it seems to be an extension of the Illuminati.

While the page might look somewhat like the Nick Riewoldt page on Wikipedia, it has one very important difference: each article has its own single author. One thing that I have learned from watching the various minor recent scandals involving pages on Wikipedia (John Seigenthaler, podcasting etc) is that reference publications have a problem if they don't get everything completely accurate all the time, especially where it involves facts about an individual person. Instead of the individual author being blamed, the publication itself is blamed, because it is assumed that the publisher takes responsibility for the content. This is especially true of Wikipedia since its articles are collaborative efforts, so that it is rare that 100% of any article can be attributed to the work of just one person. The authorship of an article on that site is often assumed in people's minds to be "by Wikipedia" or "what Wikipedia says". Contrast that to blogs: you don't assume that when someone writes a nasty article in a blog that it's Google in the wrong. Given that Tinfinger is attacking this problem right at the pointy end, with every article being about a specific person, I don't think it's sustainable to hope that the community of contributors can get every detail about every person right. Instead, Tinfinger's articles will all be assumed to be opinion from the outset. For the functionality of this system I will be borrowing heavily from blah, the software coded by WakkaWiki author Hendrik "morn" Mans to run PlanetCrap.

To that end, the articles are given labels according to their aim. At the moment, I have this list of article types:
  • General - for any otherwise unclassifiable article
  • Review - a review of a current work by the subject in question (e.g. book, movie, blog), or his/her performance at an event (e.g. sporting event, speech)
  • Interview - text or podcast, usually with the subject but sometimes about someone else connected to the subject (particularly in the case of a dead subject)
  • Encounter - a meeting between the author and the subject
  • Analysis - a broad category for any balanced essay, especially one which uses figures
  • Comparison - with one or more other people in the Tinfinger database (the article appears on the pages of all people included in the comparison)
  • Argument - an argumentative essay making a point about some aspect of the subject's life
  • Praise
  • Criticism
  • Humour - these last three are self-explanatory
  • Memories - some crossover with Encounter and Review here, more focused on older works and events
  • History - this is more like a full Wikipedia profile.
Can anyone think of anything I've missed?


Anonymous Anonymous said...

Looks nice, Paul.

Just curious, how many records do you plan to have on day 1 and how are you going to populate the database (ie. manually, from wikipedia, etc)?

Also how will you get the blogs' links (bottom right)?

8:20 am, January 03, 2006  
Blogger Paul Montgomery said...

Mario, we've been testing with people and news sources from AFL, NFL and cricket categories. We have 4,447 people and 78 sites at the moment. As you can see, we're going to have quite a huge list when we're up and running. I wouldn't be surprised to see a million people in our database in the first year.

Populating the database is all manual, sourced from anywhere I can find, including Wikipedia. For instance, the NFL list was sourced a little from Wikipedia but mostly from

As for blogs, users will submit their OPML list, bloggers will submit their own RSS/Atom feeds, and anyone can submit their favourite feeds and categorise them.

4:58 pm, January 03, 2006  

Post a Comment

<< Home