Base64 is bulky!

[ghost]

You will be able to get much more data stuffed into facebook chat if you ditch Base64 for Unicode, we also started with Base64, but IIRC it was a ~30% inflation of data size.

Check out this project I was a part of in 2009, we used Twitter as file storage system, so, pretty similar.
https://www.youtube.com/watch?v=kUcLbmU5fHk

Love the project, good luck!

[matiasinsaurralde]

Thanks for the info, I'll take a look!

[ghost]

The basic idea is you take your data blob, and encode it directly as unicode characters.

I uploaded some old proof of concept code, if you're interested.
There's a python version and C++ version of the datablob -> UTF8 encoding function, but you may have to dig around for the good stuff.
https://github.com/C-P-949/TwatFS