01:28:04 <stepkut> should I call this library ABNF or abnf
01:28:43 <stepkut> I think I prefer lowercase package names even if it is named after an uppercase initialism
01:28:45 <carter> whats it do?
01:28:49 <carter> do lowercase
01:28:54 <carter> i always wtf when thres both
01:31:42 <stepkut> parses, generates, pretty prints, etc, ABNF
01:32:08 <stepkut> http://en.wikipedia.org/wiki/Augmented_Backus%E2%80%93Naur_Form
08:15:15 <bergmark> stepkut: lowercase!
15:07:30 <indigo> Hm, I have no idea why Snap is more popular than Happstack. It doesn't even have type-safe routes by default. :(
15:08:50 <stepcut> At a time, it attracted a lot of attention because it used enumerators and had 'new project energy'
15:09:01 <stepcut> of course, now they are ditching enumerators :)
15:09:42 <stepcut> also, people get excited about snaplets
15:10:22 <stepcut> plus the code is pretty to look at :)
15:10:51 <stepcut> really, I need to get the new happstack backend out, ASAP
15:10:53 <indigo> Well, any Haskell code is pretty to look at when you're coming home from PHP work like me. :P
15:11:11 <indigo> Oh yeah, Happstack8 with pipes?
15:11:17 <stepcut> yeah
15:11:26 <stepcut> the code is in progress at last
15:11:33 <stepcut> there was a very long wait for pipes to stabilize
15:11:42 <indigo> Nice. Going to improve on those benchmarks vs. warp? ;)
15:11:53 <stepcut> absolutely
15:12:13 <stepcut> in fact, the test suite already has built in tests for benchmarking
15:12:25 <stepcut> to ensure that we don't actually commit a patch that slows things down
15:12:40 <indigo> Sounds awesome.
15:13:24 <stepcut> I am currently working on the 'reference' backend -- in which proven correctness is the primary important factor. Once that is written we can then work on a super fast backend
15:13:44 <stepcut> because there is a reference backend, we can then write quickcheck tests to test that the hand optimized version didn't screw something up
15:14:11 <indigo> Well, looks like you guys definitely have your priorities in order. Correctness first, then speed.
15:14:15 <stepcut> yeah
15:14:46 <stepcut> once the new backend is done, I will have rewritten every line of code in happstack since taking over maintainership :-/
15:15:06 <stepcut> except for acid-state/safecopy -- Lemmih rewrote those :)
15:15:43 <indigo> Damn. Sounds like a huge undertaking for a single person.
15:16:10 <stepcut> well, I've been then maintainer for 5 years now
15:17:01 <indigo> Oh, I guess it's not *too* bad then.
15:17:34 <stepcut> yeah -- next step is to ditch Haskell and switch to Idris :)
15:20:23 <indigo> Hm, I should look into Idris.
15:26:47 <indigo> Looks interesting. Dependent types and eager evaluation.
15:35:09 <stepcut> yeah
15:35:17 <stepcut> I am porting it to the arduino