Vafiadis: Tcsc_: I doubt it would help much for the frame stuff though, it’s more useful for complex data operations
Paulhus: Vafiadis: here a reference of tools or dialogues u can launch via console. u may find some of them useful . http://bit.ly/Paulhus-dos-windows-commands i did
Vafiadis: Paulhus: hehe thanks, I’ll keep a copy for next time I’m working in windows-land
Vafiadis: Tcsc_: http://jlongster.com/Transducers.js-Round-2-with-Benchmarks – “baseline” are the manual/”for loop” implemenatations
Mickelson: Vafiadis: what i mean by profiler noise is that seqiterator1, 2, 3, composemapx = x + 1, filterx = x % 2 === 0 is not going to give you good information when you look at the profiler
Nishio: The transducer will seduce ya!
Vafiadis: Tcsc_: yeah, I can see that being an issue
Vafiadis: Any kind of lazy eval is likely to give you that issue of course
Vafiadis: I mean, you’d be able to see if an inordinate amount of time is being spent in whatever function finally evaluates it
Vafiadis: But beyond that it’d be tricky
Figurski: It depends on the view you’re in. you’re basically making the bottom up view less useful
Kalinoski: What does the word “lazy” mean viz. programming
Kalinoski: Sorry I can google this
Brinker: And at that point you’ll probably have to rewrite to be the optimized way anyway
Vafiadis: Right. it’s very use-case-oriented
Struck: Kalinoski: not computing until it’s needed
Vafiadis: Overall, I’d rather get mostly decent perf in exchange for simple, non-mutating code and then optimize where needed
Vafiadis: But in some cases, yeah, the optimized way will be necessary
Lethco: Vafiadis: i can understand that, but it’s kind of like, where are you going to spend the performance? you have 10ms/frame-ish to do as much as you can, and the more you do the better the result. if you optimize one part, that allows you to spend that performance on something else, more particles or whatever
Empasis: So i forgot to git push origin develop and hopped on a diff. computer and ran git checkout -B newbranch from develop. i continue working in newbranch until i git merge back into develop right?
Kalinoski: I love discussion about performance-critical code
Mahnke: You can spend them on convenience, and it’s not bad to do so, but you only really want to do that if you get a lot out of it
Bundley: Also the really complex code tends to not benefit so much from that. i mean writing a convex hull function is moderately tricky, whether or not you use mutations
Ollivier: Opiates: sounds okay to me. there’s a git channel btw
Sherief: Hylle: Your tweets about Quebec made me sad. :/
Baillet: Deltab thanks wasnt aware
Vafiadis: Tcsc_: a lot of it depends on what your definition of “bottleneck” is going to be
Vafiadis: Tcsc_: if you’re doing a bunch of async stuff, or calculations between dom updates or something, it’s going to be pretty negligible. If you’re gonna do frame-sensitive stuff.
Pizzaro: Oh, yeah, this only applies to code that runs per-frame
Vafiadis: Yeah, if I did that more I’d probably do things more your way
Vafiadis: Lazy eval is also a PITA for really time-sensitive stuff like that because it’s hard to know when something has been evaluated
Vafiadis: I’m looking at you, Haskell
Milsaps: Yeah, that sounds tough.
Francom: Hylle: Because I think it’s a bit too easy to find something “silly” without knowing a bit more about its history
Wojtanik: Hylle: If you care a little bit: http://minorityrights.org/minorities/french-canadians/
Zamudio: That should give you a better idea.
Groman: I just find any kind of nationalism silly tbqh
Bhatti: Jtreminio: Perhaps you could trim it on “blur”
Klaman: I mean, I don’t personally see that as a bug. but :/