"J. P. Gilliver (John)" wrote

| You're right - I'd forgotten I have an indication (both graphical and
| audible) (both directions) of net traffic, and that's silent, so it
| isn't mining. I didn't think it was, really. (Though would mining
| involve a lot of net traffic, or just a bit to fetch and a bit to return
| the results?)

It seems to be flexible. I imagine something could download
a chunk to chew on, but micro-mining is being done on
webpages by just running client-side script while you're
visiting. That's even been talked about as a possible income
model for commercial sites. So it must be feasible to do tiny
bits efficiently.

| They're different purposes: Everything works on filenames, Agent Ransack
| on file contents.

AR is both. I got it because it *can* search file content.
But it also does file name search very fast. I use both
functions. But I just don't have as many files as most
people seem to. And I organize them. So I don't need
to search so much for file names.

| I don't think Everything - because of the way it
| works, which I don't understand, but it's something to do with NTFS, I
| think - _does_ hammer the disc. I use it mostly - but not exclusively -
| when I want to ask myself "have I already downloaded a file with x
| [often a serial number] in its name".

I'm surprised it needs to index if it doesn't do
content search. Can it actually be faster to search
its own list than to search the file system?