I think Google uses the 'did you mean' language. Many people, including
myself, find it awfully useful. You don't, Bernhard?
But, yeah, the key, I think, is not in the formulation of the language
introducing the suggestions, but the quality of the suggestions.
Google's work because they are actually really good. Which is not an
easy thing to do. We may or may not be able to pull it off. If we can't
pull off suggestions that actually _are_ useful, probably better not to
have them at all. But I don't think the language used to introduce them
is of utmost importance.
In Xerxes we have spelling 'corrections' only, suggestions that are just
a spell check against a standard American dictionary. It works out to be
pretty useful I think, people appreciate being notified when they've
potentially entered a mis-spelling.
Jonathan
Bernhard Eversberg wrote:
> > We currently do a comparison match called "do you mean?"
>
> Frankly, I'm appalled! I don't want to have a machine guessing
> what I mean and then address me, as a person, with that kind
> of unsolicited questions!
>
> What I mean is my business and mine alone. When I make mistakes
> entering search terms, that's my problem and mine to solve,
> using my own intelligence and judgement when looking at inadequate
> results or "zero hits".
> This kind of features only raises unrealistic expectations and
> creates a trust where mistrust in technology is vital
> and trust in one's own capacities must not be eroded but supported.
>
> This bending over backwards to spare the user the "zero hits"
> experience is, I think, counterproductive in an insidious way.
> All to easily, timid souls are apt to turn over their judgement
> to the machine instead of working on it.
>
> IOW, use neutral, unobtrusive, unpatronizing formulations for
> functions that a reader might or might not want to use, esp. when
> what the algorithm blurts out can easily be patent malarkey. Don't
> create the impression that there's more intelligence inside the
> machinery than there can possibly be.
>
> B. Eversberg
>
Received on Fri Feb 20 2009 - 10:19:10 EST