C[omp]ute

From: "andrew cooke" <andrew@...>

Date: Wed, 24 Dec 2008 11:05:23 -0300 (CLST)

While at the AGU conference I noticed this magazine (Computing in Science
and Engineering) - http://cise.aip.org/

It's not the most professional looking magazine in the world, but it is by
far the closest I've found to a magazine that covers the kind of work I
do.  So I subscribed - got a cheap price and the last 10 years on CD.

I copied the back issues from CD to my hard drive and then wondered what
to do.  They are broken down into one article per file, but they are PDF
and I had no way of searching them.

Then I remembered Beagle, which I have carefully de-installed from my
computer for several years now because it seemed (1) useless and (2) used
up so much CPU.  Beagle is the default Linux document indexing system
(think Google for your local disk).

Anyway, it's improved hugely.  It no longer locks up the computer when
you're trying to get something done.  And it appears to return moderately
useful results.  So I can now search for something like "python numerical
vector" and the second hit (when ordered by relevance) is "Python for
Scientific Computing" from the magazine.

You can also restrict to certain file types (ext:pdf), exclude terms
(-foo), etc etc.  It even indexes music metadata (so you can search for
artists) and email.

The KDE Kerry interface does quite a good job of presenting the data too -
you can click on a file and it is handed off to an appropriate handler
(although HTML is going to konquerer instead of the mozilla configured for
the desktop).

It does, however, still seem to crash quite regularly.  You can see this
by doing

  tail ~/.beagle/Log/current-Beagle

which will show a message about SIGSEGV or by typing beagle-status which
will say it cannot connect.  To restart just type "beagled" at the command
line.

I also enabled extended filesystem attributes by modifying the fstab from
/home to include the user_xattr option.  Apparently this makes things a
bit faster.

One check I did, when I found some old files I thought I had been deleted,
surprised me: it immediately picks up on deleted files and excludes them
from the results.  Which is nice.

Andrew