Hi,
On 7/2/07, Randall R Schulz <[EMAIL PROTECTED]> wrote:
> * Both Beagle and GDL present results sorted by date, not relevance.
Google Desktop has both by-date and by-relevance display options.
Ah, I missed that. Indeed it does, but so does Beagle.
In either case, relevancy in desktop search is tough to do and
probably doesn't work all that well in practice. (Part of what makes
web search easy, and what Google picked up on from the start, are that
the relationships between web pages are part of what make them
important. Sadly, those links simply don't exist in desktop data
today.)
Odd. What I've noticed over the few hours since it's been running is
that it has been entirely unobtrusive.
It's not that surprising. From personal experience and reading
various people's posts on this list about Beagle, it seems to work
very well for some and not so well for others. I think it depends a
lot on the volume and type of data.
It's now been running 4 hours and has indexed 37.4% of the files in the
directories I've configured. That's currently ~17,800 emails and ~6,000
files. I've been working away in IDEA (Java IDE) and doing ANTLR and
Ant builds along the way.
As a counterpoint, I let it run on my home directory for about 30
hours, and it reporting being a little bit more than 30% done. In
those 30 hours, the time estimate dropped from 5.8 hours to 4.1 hours.
I then left and let it run over the weekend. When I came back in
this morning my machine was nearly unresponsive. When I got the index
status page loaded, it said that was 15% done (with 5 hours to go!)
but had jumped from about 300,000 items indexed to 660,000.
The estimation stuff needs some work -- we tried to do it in Beagle
but without thrashing the disk walking every directory, email, etc.
and counting them there's no accurate way to do it. We just decided
to punt it.
Joe
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
