Dictionary Game

[iluvsheep]

I wrote an entertaining little game where you try to guess words based on three-letter substrings of those words.
Link to the dictionary game
I originally wrote this game in May 2006, but wasn't satisfied with it and lost interest. Reading over the four-year-old python code was quite an adventure. I was struck by 1) how much python I've forgotten

Comments

[archonsengine]

Oh, man, this looks addicting--partly because I've been doing so many crosswords lately, but still!

I went with panoply the first time and got substrings valued from 4 (eyt) to 723 (spi). For some of the tougher ones, I would type ridiculous guesses and find myself getting closer and closer. Two examples:

ilb (70)
failbook
failboat
sailboat

etb (30)
petball
feetball
basketball

But, yes, awesome game, thanks!

[archonsengine]

Oh, and a scoring system/scoreboard would be cool, if you decide the game needs more features. :)

[iluvsheep]

Glad you like it.

If I'm stuck I'll type in the three letters I am going for and then semi-randomly add stuff at the beginning or the end until (hopefully) I'm reminded of a matching word.

A scoreboard would be sweet - perhaps even setting a cookie and keeping track of life-time statistics. Perhaps sometime after the holidays if there's enough interest.

[archonsengine]

That's a good idea, because I keep wanting to comment again with silly or interesting things that happen in the game. I'm sure that will drive you insane sooner or later. ;)

Oh, and what dictionary are you using? I was surprised that l[yca]nthrope wasn't a word, though l[yca]nthropy was. Also, [bic]urious is apparently unacceptable (hooray for [bic]uspids!).

[paperclippy]

I like it, but I think the biggest problem with it is that the difficulty is based on how many words have that substring rather than how common the words with the substring are. For example, "afe" had something like 101. However, it's a really easy one because "safe" is a very common word. "eos" on the other hand had more words, but almost all of them are obscure scientific terms (eosin, eosinophil, eosinophilia).

[iluvsheep]

Absolutely agreed. I don't suppose you happen to know of some sort of standard "commonality" rating for words?

Even the current rankings are a little weird in that verbs usually have a bunch of conjugations whereas nouns usually only have two forms (singular and plural). I'll probably solve this additional problem with some sort of "stemming" algorithm, which my NLP friends say is well-known and standardized, even if I hadn't heard of it.

[paperclippy]

I don't know of any standard rating for words but I bet one exists. Maybe whatever system they use for picking words for spelling bees?

[archonsengine]

Alternately, for higher difficulty, you could just kick out words under a certain length.

[derakon]

I made a leaderboard system for a game I wrote, in Python. It's not too complicated; just have a script on your server that accepts the score infromation and a salted hashed string that you can use for verification.

Of course, this might be trickier since your program is partially browser-based; it becomes harder to trust input.

[dragonmudd]

I love it.

For potential feedback, I don't like that it counts the plural of a word as another solution. Of course, if you had a clever way to eliminate that, then you have to make sure that instances where the plural is a solution but the singular is not a solution don't get eliminated.

That being said: teq=2... easiest ever!