Vocabulary count stat

flav0rflav

Go on, abuse that database and giv'us a stat showing the number of unique words each poster has. It might increase the level of english of us all. Or increase the amount of alternative spellings.

p.pete

That's a cool suggestion - suppose it would have to be based on words that match up to a dictionary definition. It would probably be a bitch to implement though.

Sir Digby Chicken Caesar

good god why?

oh, and "alternative spellings"

you mean like l33t?

ecksor

Haha, that sounds funny. It wouldn't actually be that hard to do, it'd probably just murder the database ...

Hm ....

elexes

prob cause more probloms then worth

p.pete

Might be easier to occasionaly release the statistics and store them somewhere rather than keeping a running total.

flav0rflav

Come on, I'm not great at SQL. I can do it in unix, use sed to replace all white space and punctuation with newline, to get all words on seperate line. Then use sort piped to uniq piped to wc, to sort, remove duplicates and count.

p.pete

Originally posted by flav0rflav
Come on, I'm not great at SQL. I can do it in unix, use sed to replace all white space and punctuation with newline, to get all words on seperate line. Then use sort piped to uniq piped to wc, to sort, remove duplicates and count.

Yes, but do you want this procedure being run for every single post - I don't know how much you know about it but it could cause a lot of unecessary processing.

davej

How about stats on more high level things like:

which forum(s) a user posts most to.
how many threads they actually initiate (as opposed to reply to).


Based on these type of stats you could also work out which users have the most "interests" in common with each other. It could help you to find your boards.ie soulmate (or arch enemy)!

davej