Spell Up: A Chrome Experiment
11–20 of 20 posts
Re: Spell Up: A Chrome Experiment
#12So apparently Google wants to mine data for spellings? Great idea to make a game for that.
-GOOG-411 (voice recognition)
-Ingress (location data)
-Google Image Labeler (image identification)
-Recaptcha (book / streetview ocr)
I'm sure there's more, but those are the ones off the top of my head.
EDIT: They obviously mine data from everything, but these are examples of overtly creating a tool/game to collect or improve a specific type of data.
Re: Spell Up: A Chrome Experiment
#13Re: Spell Up: A Chrome Experiment
#14Re: Spell Up: A Chrome Experiment
#15The game worked for me. But the voice recognition of individual letters was abysmally slow / bad on my laptop. I guess it gives a better sense of how far back voice recognition is.
I haven't tried the system under discussion to know if it has such a mode or not.
Re: Spell Up: A Chrome Experiment
#16Re: Spell Up: A Chrome Experiment
#17The game worked for me. But the voice recognition of individual letters was abysmally slow / bad on my laptop. I guess it gives a better sense of how far back voice recognition is.
Out of 10 tries for "G" I became angry and closed it. Also a laptop.
Re: Spell Up: A Chrome Experiment
#18Good idea, but rough on the execution.
Re: Spell Up: A Chrome Experiment
#19The algorithm to convert my spoken letter to the letter that the game thinks I am guessing is wrong often enough to be annoying.
Lastly, it's too slow: I have to pause for several seconds between each guess.
Re: Spell Up: A Chrome Experiment
#20Game asked me to spell "khob". I didn't know what that was, and tried my best. Turned out it was asking for COBRA. Well. Hopefully that data was useful for something.
At the "T", I said "T" and it first heard it as "K", and next heard it as "P" (I had a retry bought).
To be fair, though, I was in a somewhat noisy room. Still, though, losing for something that's not my fault is a pretty frustrating experience.