Comments on: Google autocomplete now fixes spelling problems http://www.malcolmcoles.co.uk/blog/autocomplete-fixing-spelling/ Where to find Malcolm Coles, reviews, and tips on how to do things I couldn't do. Mon, 25 Mar 2013 16:22:31 +0000 hourly 1 http://wordpress.org/?v=3.0.4 By: Nichola Stott http://www.malcolmcoles.co.uk/blog/autocomplete-fixing-spelling/#comment-7322 Nichola Stott Mon, 18 Jan 2010 10:59:15 +0000 http://www.malcolmcoles.co.uk/blog/?p=3853#comment-7322 I agree this is a clear usability improvement on the query input side; though it would be nice if there was different handling of my request according to the selection method. E.g. if I start to type "wether" (male goat/sheep) and select "wether" from the auto-complete options - we could maybe infer I'm less likely to be interested in BBC Weather in position 1, as I have made a comparative selection - kind of pre-qualified my intent? Maybe this is on the cards and maybe all of these recent shennanigans are part of a CTR data collection exercise? I agree this is a clear usability improvement on the query input side; though it would be nice if there was different handling of my request according to the selection method. E.g. if I start to type "wether" (male goat/sheep) and select "wether" from the auto-complete options - we could maybe infer I'm less likely to be interested in BBC Weather in position 1, as I have made a comparative selection - kind of pre-qualified my intent? Maybe this is on the cards and maybe all of these recent shennanigans are part of a CTR data collection exercise?

]]> By: Malcolm Coles http://www.malcolmcoles.co.uk/blog/autocomplete-fixing-spelling/#comment-7319 Malcolm Coles Mon, 18 Jan 2010 10:25:35 +0000 http://www.malcolmcoles.co.uk/blog/?p=3853#comment-7319 It handles stationary / stationery well too - giving a mix of terms relating to both spellings if you type in stationary (as opposed to just returning stationery results if you do search on stationary). The stationery/stationary thing makes me wonder if it's CTR plus next search data. Presumably in the old days, people would search for stationary, see things that didn't move and then search again for stationery when they realised it was a mis-spelling (possibly prompted by a "did you mean"). It handles stationary / stationery well too - giving a mix of terms relating to both spellings if you type in stationary (as opposed to just returning stationery results if you do search on stationary).

The stationery/stationary thing makes me wonder if it's CTR plus next search data. Presumably in the old days, people would search for stationary, see things that didn't move and then search again for stationery when they realised it was a mis-spelling (possibly prompted by a "did you mean").

]]> By: Joanna Butler http://www.malcolmcoles.co.uk/blog/autocomplete-fixing-spelling/#comment-7317 Joanna Butler Mon, 18 Jan 2010 09:40:40 +0000 http://www.malcolmcoles.co.uk/blog/?p=3853#comment-7317 Agreed - much better user experience. Actually useful in most cases. It seems fine with US/UK spellings (e.g. favourite/favorite, colour/color, humor/humour) but why not acknowledge optimisation? Odd... If it's based on search volume then we should also be pushed 'humor' over 'humour'. But we aren't. I'm guessing CTR based. Agreed - much better user experience. Actually useful in most cases. It seems fine with US/UK spellings (e.g. favourite/favorite, colour/color, humor/humour) but why not acknowledge optimisation? Odd... If it's based on search volume then we should also be pushed 'humor' over 'humour'. But we aren't. I'm guessing CTR based.

]]>