An unofficial blog that watches Google's attempts to move your operating system online since 2005. Not affiliated with Google.

Send your tips to gostips@gmail.com.

May 29, 2009

Google's Context-Sensitive Spell Checker

Spell checkers aren't usually very smart: they highlight words that aren't in a dictionary and suggest a list of similar words. Even if they take into account words that aren't included in dictionaries and they deal with plurals and verb tenses, most spell checkers can't find words that are used incorrectly in a context.

Wikipedia includes as an example: "Their coming too sea if its reel", a phrase that has 5 spelling mistakes, even though all the words can be found in the dictionary. If you enter this text in Gmail's editor and click on "Check spelling", Gmail won't find any error. Type the same text in Google's search box, and you'll get a "did you mean" message that suggests to search for "Their coming to see if its real". As you can see, Google's search box has a better spell checker than Gmail since it doesn't rely on a dictionary, it uses a huge amount of searches to determine what are the most probable sequences of words that follow a certain pattern. Unfortunately, the spell checker available at Google.com is optimized for searches, which are usually short, so you can't use it to spell check an email message or a blog post.


Google Wave, the service demoed yesterday at Google I/O, includes a context-sensitive spell checker that highlights errors as you type. Google uses the language models built for Google Translate to find words that don't belong in a certain context.


29 comments:

  1. Doesn't know "they're" from "their"??

    ReplyDelete
  2. spelly & linky would be very nice inside Gmail.
    i think searchy already exists as a 'Labs' experiment.

    ReplyDelete
  3. @Guffy - Well, that's not the only problem with "Their coming to see if its real", but actually that mistake is relatively small compared to how it changed "its reel" to "its real". That is, "their coming" is quite a valid phrase, despite the implication by the following "to" that this is not the intended meaning.

    For "its reel" -> "its real", on the other hand, the correction makes less sense than the original! It would be different, of course, if it had suggested "it's" for "its".

    ReplyDelete
  4. @Anonymous - HUH? In the context of the sentence given, how is "their coming" considered valid? The only way "their coming" can be valid is if it's used in a sentence where "their" is actually a possessive pronoun. For example, "Their arrival was unexpected."

    The sentence shown in this blog entry would not be valid if "their coming" was considered valid. Never-the-less you are correct when you reference "its real" as a fault.

    The correct grammar would be "They're coming to see if it's real."

    ReplyDelete
  5. My example compared Gmail's spell checker with Google Search's "did you mean" message, not with Google Wave's context-sensitive spell checker. We'll have to wait and see if the new spell checker delivers better results.

    ReplyDelete
  6. I don't see why they couldn't import this new context spell checker into Gmail.

    ReplyDelete
  7. @John Yup, I didn't mean that "their coming" was valid in context the "complete" context of the sentence, as I said -- merely that "their coming" is, taken by itself, a usable combination of two words, whereas "its real" is not (at least not in any context I can think of).

    ReplyDelete
  8. This needs to be on YouTube comments. Now.

    ReplyDelete
  9. You can try Ghotit.com (www.Ghotit.com) a great Context-Sensitive Spell Checker that includes a plug in for Microsoft Word

    ReplyDelete
  10. I see people get they're/there/their incorrect more often than not. If they're basing their spell checker on popular usage, I'm not sure I want it.

    ReplyDelete
  11. Ginger is a much better context-sensitive spell checker - www.gingersoftware.com. It corrects full texts automatically and it is free...

    ReplyDelete
  12. Ginger seems to be a great software, but there are some things that could be improved:

    * it only works as an add-on for Microsoft Word, so you can't use it in a browser or in other operating system than Windows
    * you need to install an application and create an account (for free). The online interface seems to be limited to very short sentences.
    * the software doesn't correct your text as you type, you need to manually correct each sentence one by one.
    * it's very slow; you can wait even 10 seconds for longer texts.
    * it's limited to English.

    Let's compare it with Google software:
    * it works online
    * limited to Google Wave, but I don't see why it wouldn't be integrated in Gmail, Google Docs or even made available through an API.
    * it works in realtime, so it's fast
    * it's not limited to English

    ReplyDelete
  13. I think it's a nice little added capability. Sometimes the spell checkers are just a nuisance for that very reason.

    ReplyDelete
  14. I think this is fantastic and to me (background in AI, long ago: ) this is the most impressive feature of Google Wave!

    Does anyone know ANY product out there that does this? So smart and well? I would love to hear about them...

    For now: eagerly awaiting my invite : >

    γιδω

    ReplyDelete
  15. I think that the Ginger approach is better in handling context. If you correct while you type, you have very limited understanding of the context. Make sense to finish a sentence and then correct.
    I would think that Ginger will not be limited to MS Word or one OS, they never said so. Wait for their first release.

    ReplyDelete
  16. How do you use this service? How do you get this spell checker on Gmail? Kazoo from Ohio

    ReplyDelete
  17. It's true Ginger is free but not for long, I have seeing Ginger planned pricing for educational institutes and it is far from being free (most probably home user will pay more)

    ReplyDelete
  18. @John
    I think that this is an excellent example of how it is optimised for short queries, hopefully wave will be more comprehensive.

    ReplyDelete
  19. I hope that Google will add the Spell Checker from Google Wave to Gmail Spell Checker and Google Docs Spell Checker?

    ReplyDelete
  20. This spellchecker would be great...Well this is an amazing thing...Google has always come up with new applications and attracts a lot of people..

    ReplyDelete
  21. first of all google is king of the search engine industry but the other software also develop like google analytics,google insight,google adwords and google doc. there feature is new and useful also.

    ReplyDelete
  22. I usually type a phrase into Google instead of spell-checking it as it gives me more 'intelligent' results. Binding search results and spell-checker sounds great.

    ReplyDelete
  23. Dang....you people are picky....
    i put out thair whut i can to get my pont across. :))

    ReplyDelete
  24. Well, I am happy to have ad control option with the opportunity to give feedback. Shall see how we do. Just setting up on my new Vista, so getting used to lots of new controls. Shall see how Sifter works as an advertising platform.

    ReplyDelete
  25. I want to know why the Google spell checker doesn't pick up when I forget to capitalize the "i" in words like "I" "I'll" and "I'm" even though it will tell me when I've spelled the word "I've" wrong because I used a lowercase instead of a capital "i".

    ReplyDelete
  26. Google Wave is a good step in context spelling, but still behind even MS-Word-2007, not talking about Ghotit.

    You can see the sentence from Wiki well-corrected by Ghotit, and look at the comparison among the context spellers here:
    http://www.ghotit.com/contextspellcheckers.shtml

    Ghotit web-site has an online spelling and grammar checker as well as a plugin for MS-Word.

    ReplyDelete
  27. Google Wave sounds cool. But I'm a loyal consumer using this spell check tool. It's free and has grammar checker plus no need to install anything.

    ReplyDelete
  28. Does this make sense?
    I hope I'm not too complex...?
    (@_@)

    ReplyDelete