Guys, I wrote a similar tool few months ago: http://www.instantwordsearch.com

It does more than definr (e.g. pattern search). But it was met with indifference (except on StumbleUpon). Can you tell me where I screwed up?

May be I shouldn't have mentioned that I wrote it as a toy app? (but it IS a toy app) Picture needed? Looks ugly/too simple? Name too long? All I am able to come up with are some conspiracy theories: 'Definr creator lists Kevin Rose as friend in his Digg profile'; 'Ruby-On-Rails was mentioned' etc. But the real issue must be something else.

I think the fault is invisible to me as I created the tool. Hopefully I can learn a marketing lesson or two from this... Please help.

I just typed in 'quotidian' to both apps. (It came up this weekend.)

Definr is very fast. It looks up words faster than I type so that the list of words at the bottom updates with each new letter. With IWS, I had to pause mid-word for a half-second in order to get an updated list. By the time I got a choice of words with IWS, I had finished typing and already made a typo.

Also, the interface for Definr is much more friendly looking and simple. Definr gives you the complete definition on the main page, instead of links to other definitions. And Definr is an easy name.

Sorry, but I liked Definr much better.

I haven't played with it enough to evaluate it on functionality, but as-is, it's not pretty enough to get popular. It's sad that aesthetics matter so much for a purely functional tool, but that's just how things are.

I'm not sure it's incredibly fast, sure it uses auto-complete and ajax to load things instead of a page refresh, but it seems a similar speed to conventional web based dictionaries to me.

It updates the word list underneath the search bar with every letter typed...at full speed typing. Since typing speed pretty much defines the maximum data input speed, I think that qualifies as 'incredibly fast'.

heh I wasn't that bothered about the 'suggest' feature. You still have to wait for the definition to appear.

Still cool, but I'm not sure it's much faster (Assuming you don't care about the suggest functionality).

I was interested in the suggest feature mostly because I didn't know how to spell "quotidian"!

Otherwise, the differences are more about presentation than an actual difference in the technology. I guess I was really just impressed by the user interface of Definr.

This cries out to be implemented with Berkeley DB, which is simply much faster than relational databases.

I'm not sure it's useful either. Why not just use a dedicated data structure (such as a radix tree dictionary)?

I had completely forgotten about BDB, thanks for the reminder!

Nothing will ever be faster than hovering over a word and pressing control-command-D on Mac OS X.

Sweet. However, the one word I happened to pick for a test (don't know why) was "spork" -- doh.

At least it wasn't a presentation at YC: http://news.ycombinator.com/item?id=77269

How is this a startup? Auto-complete is NOT hard: it's a trie.

I see no indication that it's intended to be one.

I've been using this one. Easier to type definr though.

http://www.ninjawords.com/

Making a round trip with each letter typed? Impressive. Who needs a desktop app?

I tried "triumphalism" on the first try. Not yet there :)