Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
Expensive, difficult, slightly dangerous, not particularly effective... I'm on a roll.

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.



Spellchecj by tupo

Prioritise spelling suggestions based on key proximity
  (+5, -1)
(+5, -1)
  [vote for,

Consider this typo:


Spelling correction suggestions could be "Inch, Inca, Since, Nice, Once, etc..."

I propose a keyboard-proximity filter (based on the system's current designated keyboard layout) to order them in the most likely typo order. O is very close to I on the keyboard, so:

Since, Nice, Once are probably the most likely words.

Inca and Inch are much less likely to be typed by accident, and accidents form 90%+ of my spelling errors.

Add to that theme with some intra-sentence grammar checking and common word tagging, and spellchecking could be much more useful.

(How often does one really intend to use the words "tot he" in a sentence? it is more likely to be "to the", for example. Some systems auto-correct that automatically, though.)

not_only_but_also, Aug 17 2009

US Patent 6,801,190 http://www.google.c...AAAAEBAJ&dq=6801190
[jutta, Aug 17 2009]

Context-sensitive spell check in Microsoft Office 2007 http://blogs.msdn.c...6/06/05/617653.aspx
[jutta, Aug 17 2009]

Context-sensitive spell check in Google Wave http://googlesystem...-spell-checker.html
[jutta, Aug 17 2009]

Wikipedia: Damerau-Levenshtein distance http://en.wikipedia...evenshtein_distance
Edit distance with bells on. [jutta, Aug 17 2009]

Wikipedia: Needleman-Wunsch algorithm http://en.wikipedia...an-Wunsch_algorithm
This very clearly needs to be worked into a popular "dance craze" song, "Do the Levenshtein-Damerau Needleman-Wunsch". [jutta, Aug 17 2009]

Please log in.
If you're not logged in, you can see what this page looks like, but you will not be able to add anything.
Short name, e.g., Bob's Coffee
Destination URL. E.g., https://www.coffee.com/
Description (displayed with the short name and URL.)

       All good ideas, and patented and implemented in a few systems. A patent- or literature-search for "spell checking algorithms" might be in order.   

       The keyboard proximity thing is implemented, if one bothers with it, as a "confusion matrix" that, given two keys, tells you how likely they are to be confused. When computing the edit distance between two words (-> Levenshtein distance), instead of assigning equal probability for each substitution error, the confusion matrix is used to look up the possibility of this specific error.
jutta, Aug 17 2009

       It seems intuitive.
normzone, Aug 17 2009

       //Consider this typo:   


       Well, duh... plainly obvious you mean Vince
vincevincevince, Aug 17 2009

       //Levenshtein// - so that's what it's called - I once wrote a program that was intended to act as an "engine" for ALL card games, from snap through Gin Rummy to any/all variants of Poker, with each ruleset defined as a (relatively easy to edit) xml file - the tricky part came during draw/replace scenarios, trying to get the machine to try to decide whether it had a good/bad enough hand to draw a card (and decide which one to burn in the process), and there are lots of routines that reference the Hamming distance between a given hand, and a target one (e.g. four of a kind, or a series of hearts, or a numeric sequence) that the program might have "wanted" - I'm now going to have to go back and rename some of my methods to usd the word "Levenshtein".
zen_tom, Aug 17 2009

       //confusion matrix// - I'm pretty sure I can implement that myself without any algorithms.
wagster, Aug 17 2009

       I no. Pathetik isn't it.
wagster, Aug 17 2009

       //I think spellcheckers should be programmed to deliberately fail every so many words, or even insert barely- noticeable typos whilst typing that won't show up on the finished-product spellcheck.//   

       I find that happens already, as some errors form another word.   


       (The first to are quite common on the net, and widely reviled.)
Loris, Aug 18 2009

       One typo I often come across is a 'dyslexic' (no offense to dyslexic people) error - hitting the (theoretically) correct key with the wrong hand (eg. putting 'k' when you needed 'd').
<Pet peeve> People getting 'than' and 'then' mixed up! Grrr!</pp>
neutrinos_shadow, Aug 18 2009

       //I can't believe editors, who used to have to earn their pay by proofreading, can cheat...//   

       Yeah!! And what about those lazy sailors who use GPS to navigate..?
shudderprose, Aug 19 2009

       //The first to are quite common//   

       Was that "to" intentional? Yeah. Must've been.
theleopard, Aug 19 2009

       I know someone who frequently misuses "of" - as in must of, could of, should of etc - I don't have the heart to tell them.
zen_tom, Aug 19 2009

       You must learn to Give In To Your Hate, [zen].
8th of 7, Aug 19 2009

       //Was that "to" intentional? Yeah. Must've been.//   

       It was now.
Loris, Aug 19 2009


back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle