Re: [MORG] WGLC on draft-ietf-morg-fuzzy-search-02; respond by 10 Feb 2010

[Barry Leiba <[e-mail]>]

> This makes two (in my mind) false assumptions.
>
> The first is the assumption that the majority of servers will offer both
> substring and fuzzy searching; or even that the majority of servers that
> have fuzzy searching will offer substring searching.
>
> The second is that clients desire to control this.

I don't think anyone thinks clients want to control it; they think
users do.  This user, at least, would certainly like to.  There are
many times when I *do* want to disable the fuzzy search or the
word-bounded search that the server does today.  Very, very much.

But I absolutely agree with Mark's characterization of the first
assumption.  The commercial IMAP server I'm most familiar with, IBM's
Lotus Domino server, does its own style of search that we'd definitely
call "fuzzy".  If you search for "walked", you'll also get matches for
"walk", "walks", and "walking".  And if you search for "tall", you'll
match "taller" but not "tally".  I can not imagine that they will ever
change that, nor that they might offer an alternative.  [I couldn't
speak for them when I worked there, and I certainly can't now, but I
think this is a very safe bet.]

Similarly, I'd be surprised if Google should offer a strict substring
search as an option.  Do we think they plan to?

And, so, however much users might like the option, do we *really*
think they'll get it in practice?

Barry