...zee> Also, open-source OCR sucks because there is no open source corpus of data for training it.
May 01 08:05:58 <sonia> I've never taken one of those; very fond of road-tripping around Northland though
1.58 MB (215,511 words) - 23:33, 28 January 2015
[09:49] <Dcoetzee> I wouldn't abuse your private data :-P
[11:34] <Mono> Dcoetzee: my data plan costs 2x more than my phone
1.71 MB (227,625 words) - 19:00, 8 February 2015
�01[22:18] Zombie_Jesus using zelazny.freenode.net Corvallis, OR, USA
[05:50] <foks> Good morning USA!
1.63 MB (214,268 words) - 16:57, 8 February 2015