dons

10 ms·

91.
by dons 13y ago
I thought a lot of this stuff was already running on OKL4. E.g. https://code.google.com/p/docl4amss/wiki/OkL4
92.
by dons 13y ago
There's a 2 minute delay on the feed in case they want to [redacted]
93.
by dons 13y ago
> pure FP with Haskell or Erlang, but staffing becomes too much of an issue to scale that out Does it? We've hired several dozen Haskell developers at this point, and we always get far more applicants than available positions.
94.
by dons 13y ago
I saw the ad. Kudos for getting onto the tube.
95.
by dons 13y ago
I.e. Gergo is compiling an eso lang to a single instruction set machine that runs in the MMU "processor" that lives between the cycles on your processor.
96.
Compiling Brainfuck to the MMU Turing tarpit via a MOVDBZ machine(gergo.erdi.hu)
6 points by dons 13y ago|1 comments
97.
by dons 13y ago
If they don't have vim or emacs open, they're not hardcore.
98.
by dons 13y ago
It's doing symbolic equivalence checking via an SMT solver, iirc.
99.
by dons 13y ago
http://www.csl.sri.com/users/rushby/slides/jaist07.pdf See "SMT Solvers: Disruptive Innovation in Theorem Proving", Rushby, 2007. > SAT solving is the quintessential NP-complete problem. But now
100.
Announcing GHC cross-compilation for iOS(reddit.com)
8 points by dons 13y ago|0 comments
101.
by dons 13y ago
GIGO. Most commercial Haskell jobs are in finance -- where you don't get to contribute code to github. That leaves the PhD students and open source folks, shifting the data sideways.
102.
Making builds purely functional: cabal sandboxes for Haskell(coldwa.st)
3 points by dons 13y ago|0 comments
103.
by dons 13y ago
Calling monadic bind "flatmap" makes me cry inside.
104.
by dons 13y ago
> The core app is 9K lines of code. I don't think 9K is "large" for an application. At all.
105.
by dons 13y ago
It is monomorphically typed, however, which limits the use of higher-order functions.
106.
by dons 13y ago
I suppose it should have gone like this: * You need to unmarshal hardware floats * You visit Hackage, and use the IEEE754 package. http://hackage.haskell.org/package/data-binary-ieee754 If that is fast enough, you stop
107.
by dons 13y ago
The internal module isn't actually hidden. toForeignPtr should get you there: http://hackage.haskell.org/packages/archive/bytestring/0.9.1...
108.
by dons 13y ago
I still don't understand why you even raised a reddit thread -- "Go with Data.Binary.IEEE754 and see if there's any reason not to like it -- it just works"
109.
by dons 13y ago
Are they preemptive?
110.
by dons 13y ago
There's lots of interesting optimizations enabled by immutable data. E.g. in generational collectors old data can never point at new data, which can be used to gain additional performance. see e.g. http://research.microsoft.
111.
Mio: A High-Performance Multicore IO Manager for GHC [pdf](haskell.cs.yale.edu)
226 points by dons 13y ago|78 comments
112.
by dons 13y ago
sum is fine in GHC. It is specialised for Integer. GHCi uses naive sum though.
113.
by dons 13y ago
Yes.
114.
by dons 13y ago
Yes, the M is for monad. Eg map vs mapM
115.
Measuring the Haskell Gap [pdf](leafpetersen.com)
148 points by dons 13y ago|35 comments
116.
by dons 13y ago
"Intel Labs Haskell Research Compiler" ... !!!
117.
by dons 13y ago
Every benchmark is different. And especially if you use lazy lists instead of loops with unboxed parameters. Still, it looks mostly like a test of the random number source. And also none of the nice high level optimizations are firing anywa
118.
The Indexed Privilege Monad in Haskell, Scala, and C#(gist.github.com)
1 points by dons 13y ago|0 comments
119.
by dons 13y ago
I strongly prefer the regular composition order. But I also use fclabels (FTW!) Changing the direction of composition makes no sense.
120.
by dons 13y ago
... And S&P closes at all time high.

More ›