IMDB redesign splices a new film history

There are a couple of comments on the redesign of a favourite Web site I'd like to make. Yes, the venerable Internet Movie Database (www.imdb.com/) redesigned its look a couple of months back. Even though I use the site dozens–sometimes hundreds–of times a month, I've been fairly indifferent to the changes. The old design (which persisted for about 600 years in Net time) was good for quick glances at movie ratings and viewer comments–perfect, in fact, for a five-second visit. The new layout often requires a page scroll or two in order to glean the same information. I initially found it annoying.

But I adjusted. Anyway, it's not as if I could do anything about it. I guess the intent was to feature more categories of information in order to promote more interaction. It's more obvious now that people can contribute their own input to the knowledge base in ways other than just writing a review or giving a star rating out of 10. The result gives a bit of a Wikipedia feel to the site, which is ironic since IMDb was the original freaking-huge user-built site on the Internet. It was Wiki a dozen years before Wiki came into usage. And it's all searchable and cross-linked, as if the Web was built for it to exist in. So this was probably less of a Wikification than a redesign to highlight some information categories that had a low profile or hadn't existed before. And maybe to prompt a few more people to add opinions and obscure facts to the mix.

For example, there was always a category for genre keywords like drama or mystery, but now there's a place for people to contribute plot keywords too. Thus, the entire canon of motion pictures is being tagged with descriptors like love triangle, revenge, rogue cop, and man versus machine. But whether it was intended or not, the general public has gone far beyond using just a small list of shorthand screenplay pitches to make up its own terms based on plot elements other than the story arc. Thousands of such terms have emerged (www.imdb.com/Sections/Keywords/). Not all will flourish or be used very often, but establishing the common terms for things is a big part of this early stage of information resource-building.

Take, say, Citizen Kane. About 90 keywords are appended to it so far. Not all are useful, interesting, or apt (estate, exotic, fence, 1870, fog), but you can see how others could provide unique paths to explore filmdom. Take business tycoon, fictional biography, politician, or ambition. There are even labels that describe attributes of the film itself: no opening credits, deep focus, flashback sequence, and character name in title. As more films get tagged, you'll be able to draw a film-school thesis out of some of these cross-genre categories.

But that's not what prompted me to mention this. It's the oddball categories some people have invented. Citizen Kane breaks down as an example here, because it doesn't really have any of those–except perhaps snow globe, a descriptor shared by 16 other films to date, including Mary Poppins and Elmo Saves Christmas. Other films (less reputable ones, perhaps) get the really weird tags. How about dancing nun, Zamboni, returning character killed off, rubber band shooting, James Bond spoof scene, tattoo on butt, group vomit, Magic 8-Ball, or moose attack? Now those are some good ideas for hosting a themed movie night.

The other thing the database is enhancing is the Recommendations section that shows up at the bottom of the page after you've searched for a movie. Unlike the plot-keywords feature, which I'd say has the potential to evolve into something quite interesting as data gets entered, the recommendations come across as very much under development. You can get along just fine without someone inventing a plot keyword like Model-T Ford, but without enough intelligent data entered, the recommendations section comes across as inaccurate and silly.

I realized that a few weeks back when I searched for a 1955 film noir called Finger Man, about an undercover informer taking down a crime boss. The recommendations were so ridiculous that I actually wrote them down. First was Ichi the Killer, a disturbing and violent Asian crime film. Yes, I can see how both films are in the same family–and I've enjoyed watching them both–but you do not want your grandmother renting Ichi simply because it came up on the page. There's a big gulf between 1950s crime films and modern ones.

The next suggestion was Scarface, but the Al Pacino version, not the more appropriate 1932 original. Then there's Johnny Firecloud, a 1975 exploitation flick with a Native American racial-revenge angle. It's starting to look like only Quentin Tarantino built this list, isn't it? Next was Beyond Re-Animator, which I haven't seen. I did see the original, though, an over-the-top horror movie with some good touches of humour and 1980s self-awareness, based on an H.P. Lovecraft story. Again, not a good choice for Grandma to rent. And, to round things out, what the hell is Arnold Schwarzenegger's Kindergarten Cop doing on this list?

However, things have improved somewhat in the few weeks since I did that search. I just tried again and came up with two actual noir-type films, The Big Heat and White Heat, plus Federal Bullets, a pre-noir FBI-style film so obscure that not even the minimum five people have given it a star rating. Then there's East Side Kids, the initial serious film that eventually led to the Bowery Boys series of comedies. It's not particularly noirish, but at least it won't kill Granny. Finally, only one film from the first list reappeared. Oddly, it's not Ichi the Killer, which could legitimately be considered a true heir of the noir tradition. It's Kindergarten Cop, which I've never seen. And the really sad thing is, now that it's come up twice, I feel I have to rent it in order to find out why.