Would it be wise to create a passive daemon that has access to a database like MySQL, which receives all autolearn emails (asynchronously) and takes its time to pass them through to SA one after the other (streamlined)?
Is autolearning really that important?
*shrug*
I consider autolearning useful, but I don't consider skipping learning on 5-50% of autolearnable emails to be enough of a problem to be worth all the extra code.
It just seems like a lot of work for a relatively small gain on a non-critical function. Particularly considering that even if you do autolearn every autolearnable email you've still not replaced the need for manual training.
