Re: [MLUG - DISCUSSION] Google Desktop annoyance

[Jonathan King <EMAIL:PROTECTED>]

On 6/10/05, Jerry Gamblin <EMAIL:PROTECTED> wrote:
> > Advance warning would have been appreciated.
> 
> Like on the download page where it says "Requires Windows XP or
> Windows 2000 SP 3+".
> 
> http://desktop.google.com/
> 
> That kind of advance warning?

Now, be nice.  MIke is what we might call an "older guy".  His
eyesight isn't as good as it used to be, and that was pretty small
print.  :-)

Meanwhile, on the topic of "desktop" search, I noticed I've had Tiger
on my notebook for over a month now, so I'm beginning to get a feeling
for what its built-in search (Spotlight) is good at.  And the
conclusion is:

    It's really pretty good, but it's much better when you're looking
for something you
    know is there.

That sounds trivial, but it's not.  What I mean by this is that
Spotlight shines (heh) when your intention is to (re)find some
document on your drive and you happen to know a  keyword combination
that is unique or close enough to it.  This is because if there are
only (say) six hits on the hard disk, you'll instantly (it's really
fast) see all of them, and you'll find the one you want really
quickly.  But if there are a total of (say) 147 hits for a search,
Spotlight will try to guess which one you want, and its heuristics,
while reasonable, don't always work the way you think.  Google has
fewer problems with this when you search the web because it can use
PageRank as a guide; your personal files don't have that kind of link
structure.

So it seems to me that people who seem to like the desktop search
tools are people who really are in "retrieval mode" rather than in
"hmm, what do I have on the topic of..." mode.  The latter is a
perfectly legitimate purpose, of course, but just not one that is as
likely to work well.

To make it concrete, a search on my computer for "foxp2 projection
neurons" gives exactly the results I want.  (If anybody's hard disk
has this on it, it should be mine. :-))  But a search for "yield
curve" is not so nice, because it lights up too many things, and none
of them are obviously relevant.  Moreover, changing this to "inverted
yield curve" doesn't do much better; many fewer hits, but the top hit
is a microscopy primer.  Not at all what I was hoping for...

jking

_______________________________________________
discussion mailing list
EMAIL:PROTECTED
http://mlug.missouri.edu/mailman/listinfo/discussion