Fork me on GitHub
#off-topic
<
2017-02-10
>
fiddlerwoaroof01:02:48

@tbaldridge there actually are open source shen implementations: http://www.shenlanguage.org/download_form.html

seahckr03:02:56

any clojurians around Seattle, WA?

Alex Miller (Clojure team)04:02:27

there’s http://seajure.github.io/ - afaik that’s still alive?

stand04:02:40

Seajure still meets on the first Thursday of the month https://www.meetup.com/Seajure/

emccue15:02:30

Does anyone know where to find a good list of companies in Boston hiring for summer internships

nickbauman16:02:13

Hey I just discovered core.matrix and I’m blown away.

nickbauman16:02:24

Am I the last to know about this?

nickbauman16:02:31

Sooooooooo amazeballs.

nickbauman16:02:27

@emccue I’ve had a couple of recruiters hit me up from time to time for a couple of places to go work in Boston specifically for Clojure work. I don’t remember what they were though (was having too much fun in my current job to consider it). Want me to go check and report back?

arnaud_bos16:02:50

Yes, core.matrix is great. I'm in the process of trying to make it self-hosted ClojureScript compatible in order to include it in a blog post. I don't have much time to put into it atm though...

nickbauman16:02:24

Oh, it only works for Clojure?

arnaud_bos16:02:56

Nope, also ClojureScript but trying to use it with klipse didn't work out of the box

nickbauman16:02:36

oh! that’s too bad.

emccue16:02:44

@nickbauman I wouldn't refuse, but for sake of clarity i should put it in context that I am fairly (3ish months of messing around) new to clojure as a whole

schmee16:02:37

nickbauman have you tried any other clojure matrix libs like vectorz-clj and neanderthal?

schmee16:02:43

just curious about how they all compare

nickbauman16:02:05

No. I just assumed everyone was using clojure wrapped around JBLAS.

nickbauman16:02:14

I’m stupid that way.

jsa-aerial16:02:39

vectorz-clj is integrated with core.matrix - this is one of the really great things about core.matrix - it is protocol based and you can plug in these different backends. neanderthal is amazing, and there is at least an initial start of having it work with core.matrix

nickbauman16:02:02

Interesting.

jsa-aerial16:02:58

one odd thing about neanderthal is the owner is not so interested in core.matrix and in helping in any way with having neanderthal work seamlessly from it

jsa-aerial17:02:14

You can read various discussions on core.matrix github and on neanderthal github on this. It isn't really clear to me.

jsa-aerial17:02:18

It sucks because it basically means you (joe/jane user) need to have two totally different code bases...

nickbauman17:02:13

Meh. I would go with core.matrix. vectorz-clj is super fast.

qqq17:02:27

I thought everyone just used Tensorflow + stochastic gradient descent. @nickbauman: what do you need a NLA lib for?

nickbauman17:02:28

(from my reading, not experience)

nickbauman17:02:01

@qqq statistical processing,

nickbauman17:02:20

I’m doing a side project with my city’s crime stats.

qqq17:02:39

There's also Incanter (R in clj) ? or is that dead?

qqq17:02:41

I felt the same way about Python until I realized "library support > language elegance"

qqq17:02:56

numpy/tensorflow beats anything Java/Clj has to offer

nickbauman17:02:26

Really? I’m using Pandas right now and I’d much rather be using Clojure for this. For no other reason than I have the assoc function available.

nickbauman17:02:16

(and friends)

qqq17:02:55

I completely agree that Clojure > HyLang, but I think (by evidence of using Pandas), we both agree that in the end, we pick based on "library support" rather than "langauge elegance"