Base64 encoding versus traditional image hosting benchmarks

That's very interesting -- I never actually thought about base64 encoding an image to send inline in a file, then reconverting it with javascript!

Mostly I love how you had an idea, then tested it. Thank you for writing up the results and publishing them, too.

This could be more interesting in the future as bandwidth increases, but latency will still be limited by the speed of light. At some point, it will be worthwhile to stuff another 100kb into a file in order to prevent the latency.

Also, I wonder how much of that +100kb in the base64 encoding would be compressed back down by mod_gzip (assuming you're compressing your http connections, which is pretty standard). If that base64 overhead can be cut in 2/3, then you'll really have an interesting idea.