Tiny now, big at the time proto ML type stuff using Federal datasets. Fancy indexes into US Code / Federal Register / Congressional record etc. The big trick was auto cross referencing and linking common phrases and citations and stuff.
The dev machine (and only server it ever had) was an ebay UltraSPARC 6k stuffed with RAM and SCSI channels. It had 1tb of disk on 30+ spindles; 1200w of PSU just for disk and a carefully sequenced spin up so as not to eat the PC power supplies i stacked up for that. Kept those singing for 6 months+ building phrase indexes.
This looks really cool. I have a couple automated control programs that move little motors and such in the real world. These programs can make great use of `compare_and_set` and `swap`!
4 comments
[ 3.6 ms ] story [ 19.6 ms ] threadThat was all I could release; the code that was using that was lovely and never got to fly like it should have, alas.
Can you describe what the "lovely code" did?
The dev machine (and only server it ever had) was an ebay UltraSPARC 6k stuffed with RAM and SCSI channels. It had 1tb of disk on 30+ spindles; 1200w of PSU just for disk and a carefully sequenced spin up so as not to eat the PC power supplies i stacked up for that. Kept those singing for 6 months+ building phrase indexes.