This is COOL! I'm scanning the web right now for all sorts of stuff.

It learns from what you KEEP and what you DISCARD. If you KEEP a picture, you'll start seeing more pictures like the one you wanted. And if you DISCARD a photo, you see less of them.

It's like having a robot browse the web for you.

I was very excited to see this first public release. I've been following your technical papers in the AI journals. So far, it's working well for me, with the following little glitches:

It can sometimes take a long time to find the first images. I realize it's doing a lot of "spidering" but a more aggressive algorithm to find images at "all costs" at the beginning may be beneficial.

It doesn't detect if my internet connection drops. And if it does, it'll ZIP through its pending list and be "done" in a matter of minutes

You also should have utilities to clear the temp files in the image temp and the database

I love the "Windows Media Player" look and feel. How did you do that? I heard you say you wrote this all in Managed C++. I guess that's how you got the performance so high.

Wow! I had about 1000 downloads the first day and many nice personal emails.

The biggest "complaints" are:

- The memory footprint is big

and

- The database it generates is big

I've reduced the size of both considerably, look for release 0012X in the next week or so.

I'm a new user of Callipygian 2.0 and my first complaint at this point is that the viewing area is too small; there should be a "full screen" option, or a screen that's a little bigger anyway.

Hmmm, this looks neat.