caseyporn

Dope or Wack?

I made a new webtoy. It's called Dope or Wack? and is a quiz about modern urban slang. It's reasonably offensive, and you need Flash, and sound is reccommended. Please, share it around, and let me know what you think.

The quiz grew out of my burning desire to use some of the information at Urban Dictionary. The site is a real goldmine; not in the sense that it is excellent, but because there are some nuggets hidden in among a whole lot of dirt.

UD.com doesn't seem to have been made with any serious lexicographical intent, and that's a constant disappointment. As a resource, the information they provide is very poor: for any word, each person submits a meaning (or meanings) and an example of use. There is no notion of sense (single meaning of a word, of which there can be many) or part of speech, and a single meaning can be repeated by multiple contributors. The only quality control is a rating system ("thumbs up" or "thumbs down"), which is an unfortunate conflation of "I like this meaning" and "this is an accurate meaning".

Also, there's no information kept about where the usage was recorded - what social or geographical community a particular usage is associated with. If user information was kept (a profile of some kind), this type of dialect data could be inferred from shared words or shared location, etc. In general, I think that it could have been a much richer resource than it is.

A stroke of genius, though, was allowing user-submitted pronunciations of words. Again, the quality varies incredibly, but there are some pearlers in there. All the voices in Dope or Wack? are taken straight from the UD submissions.

The original intention of Dope or Wack was to use a much larger pool of definitions, and randomly choose the sets to guess from. Well, that's not quite right: Each "round" of the quiz would involve matching ten words to their ten definitions, and the quiz would have three rounds. At first, the definitions in each round would be random, but as information was compiled on which words were "easier" (were correct more often), the rounds would start to become "easy", "medium", and "difficult". This is an approach I've used before (in my name compatibility web toy, nom d'amour) to judge the "goodness" of a name - all names start off equal, but the more a name is used or chosen, the more legitimate it becomes.

In the end, I opted for a manually chosen set of definitions. This meant more work on my part, and less replayability, but was necessary for quality control. Even restricting the definitions to those that are positively rated, there were lots of definitions that were either too weird, too poorly written, or just not amusing. This way, there are only ten questions, but the words to guess are (I think) at least interesting. The replayability falls as a result: now, the only fun in a replay is in finding out the other words.

The soundtrack was one part I was very happy with. The loops were culled from my collection of hiphop, and I tried to pick things that the average person wouldn't necessarily know, but probably should. Some old, some new, some Australian, some Japanese, some US, some UK. I was going to keep the names secret and run some kind of quiz/prize for guessing them all, but thought it would be nicer - and kinder to the artists - to let everyone know who was getting their heads nodding.

I had (have?) grander plans involving rhyming dictionaries, too, but that's for some future time. For now, enjoy finding out what "slarp" means.

Posted by Casey at June 25, 2004 04:49 PM