[aeflash]

Individual characters can be made out, but words are hard to read. Not enough separation between letters and symbols, so constructs like "(pred=%i)\t" are harder to visually parse. I'm also not a fan of aliased pixel fonts -- it really exacerbates the picket-fence effect while reading.

Here's what your code sample looks like in Source Code Pro:

http://i.imgur.com/uAknyGB.png

Smaller font:

http://i.imgur.com/m0p17tm.png

Notice how all letters take up the full horizontal width so words clump together. It's easy to see that a space is a space, and not just an "i" next to an "l"

It's also worth noting that I had to try 3 separate OCR processors with your sample before I got one that didn't output complete garbage.

[a-nikolaev]

Thanks for such a great analysis! You pointed out some things I did not really think about. Especially the way Source Code Pro handles "i"s and "l"s.

I will try to add 1px of horizontal space between letters. The goal was to make the font vertically compact, but still legible. The size is small and it feels quite clumped though.