One nitpick I noticed, reduce is eager, not lazy.
On Tue, Apr 15, 2014 at 9:25 AM, Alex Vzorov wrote:
> Hi, Cecil.
>
> The difference is that doseq is used to produce side-effects, which is
> idiomatic, which is exactly what you did. When using doseq you can be
> sure that it's body will evalua
Hi, Cecil.
The difference is that doseq is used to produce side-effects, which is
idiomatic, which is exactly what you did. When using doseq you can be sure
that it's body will evaluate, whereas with map, for, reduce you can't -
they expect pure functions, and produce lazy sequences, and there
On Sat, Apr 12, 2014 at 5:13 AM, Cecil Westerhof
wrote:
> But it looks a ‘little’ cumbersome. Is there a better way to do this?
Here's another way. Not the best way, but I offer it to introduce you
to atoms [1] if you're not familiar yet.
[1] http://clojure.org/atoms
(def numbers '(4 6 8 10))
2014-04-12 20:36 GMT+02:00 Fergal Byrne :
> It's always best to see how best to do something, one day it'll make all
> the difference.
>
I agree. Most people call it gold-plating. And of-course you have to be
careful about that, but things like this are good to do is my opinion.
> On Sat, Apr
Cheers Cecil,
It's always best to see how best to do something, one day it'll make all
the difference.
On Sat, Apr 12, 2014 at 7:03 PM, Cecil Westerhof wrote:
> 2014-04-12 16:18 GMT+02:00 Fergal Byrne :
>
> That's fine, but note that creating new Calendar objects has an overhead,
>> using Syste
2014-04-12 16:18 GMT+02:00 Fergal Byrne :
> That's fine, but note that creating new Calendar objects has an overhead,
> using System.currentTimeMillis() is a static OS call which your now()
> probably uses, as well as object creation. It probably won't be millisecond
> sized, but you can measure y
2014-04-12 18:19 GMT+02:00 Fergal Byrne :
>
> On Sat, Apr 12, 2014 at 4:55 PM, Cecil Westerhof
> wrote:
>
>> (->> numbers
>> (reduce timed-foo [])
>> (map format-time)
>> println)
>>
>
> (->> numbers
> (reduce timed-foo [])
> (map format-time)
> (map println))
>
Tha
On Sat, Apr 12, 2014 at 4:55 PM, Cecil Westerhof wrote:
> (->> numbers
> (reduce timed-foo [])
> (map format-time)
> println)
>
(->> numbers
(reduce timed-foo [])
(map format-time)
(map println))
--
Fergal Byrne, Brenter IT
Author, Real Machine Intelligence with
2014-04-12 16:27 GMT+02:00 Fergal Byrne :
> (defn timed-foo [times n]
> (let [start (now)]
> (foo n)
> (conj times [n (- (now) start)])))
>
> (defn format-time [[n t]]
> (format "%2d threads took %7d milliseconds" n t))
>
> (->> numbers
> (reduce timed-foo [])
> (map format-t
For storing timing results of expressions, I make use of an agent to store
them in, and a modified time macro.
(def timings (agent []))
(defmacro send-timing
"Evaluates expr and prints the time it took. Returns the value of
expr."
{:added "1.0"}
[agnt expr]
`(let [start# (System/curren
2014-04-12 16:18 GMT+02:00 Fergal Byrne :
> That's fine, but note that creating new Calendar objects has an overhead,
> using System.currentTimeMillis() is a static OS call which your now()
> probably uses, as well as object creation. It probably won't be millisecond
> sized, but you can measure y
(defn timed-foo [times n]
(let [start (now)]
(foo n)
(conj times [n (- (now) start)])))
(defn format-time [[n t]]
(format "%2d threads took %7d milliseconds" n t))
(->> numbers
(reduce timed-foo [])
(map format-time)
println)
On Sat, Apr 12, 2014 at 3:15 PM, Cecil We
Hi Cecil,
That's fine, but note that creating new Calendar objects has an overhead,
using System.currentTimeMillis() is a static OS call which your now()
probably uses, as well as object creation. It probably won't be millisecond
sized, but you can measure your (now) using criterium.bench to be su
2014-04-12 15:49 GMT+02:00 Cecil Westerhof :
> 2014-04-12 15:06 GMT+02:00 Fergal Byrne :
>
or to convert your code to something more functional (and including defns
> for now and foo):
>
>>
>> (defn now [] (. System currentTimeMillis))
>>
>
> I already defined it as:
> (defn now []
> (ne
2014-04-12 15:06 GMT+02:00 Fergal Byrne :
> For precise timing benchmarks, use criterium [1].
> For simple, gross timing, use
>
> (map #(time (foo %)) numbers)
>
That is not going to work, time prints the time instead of giving it back.
or to convert your code to something more functional (and
Hi Cecil,
For precise timing benchmarks, use criterium [1].
For simple, gross timing, use
(map #(time (foo %)) numbers)
or to convert your code to something more functional (and including defns
for now and foo):
(defn now [] (. System currentTimeMillis))
(def numbers '(4 6 8 10))
(defn fo
16 matches
Mail list logo
