Fork me on GitHub
#beginners
<
2016-04-25
>
bruceadams00:04:42

prelude is certainly very good. I’ve been enjoying https://github.com/bodil/ohai-emacs

jswart02:04:46

Spacemacs is also incredible. I went from emacs -> prelude -> spacemacs

nathansmutz02:04:09

Thanks simple_smile Any thoughts on Spacemacs in emacs mode?

jswart02:04:09

mostly because I was a vim person first. but spacemacs by itself is a truly great editor in vim or emacs mode.

jswart02:04:47

If you don’t already have an editor you would die for, then I really recommend spacemacs.

jswart02:04:00

It is incredibly well done, easy to configure + add things too, etc.

nathansmutz02:04:15

If I learn primarily Spacemacs (in emacs mode) are there any speed-bumps I'd need to be aware of if I need to pair on emacs classic?

jswart02:04:53

I don’t think so. The best thing about spacemacs is it puts a layer of use/UX in front of emacs which is very nice. At work I pair with people who use standard emacs and we have no issues. The only speedbump is when my pair misses things he has configured specifically in his emacs setup. In the case of my not being able to really use his standard emacs setup we just used spacemacs.

jswart02:04:21

I suspect much of my issue is that I used vim for many years, and still use vim bindings in spacemacs.

jswart02:04:34

so i never learned the emacs bindings, and don’t personally like them enough to try

jswart02:04:33

At this point 5 or so of us use spacemacs so its becoming the most standardized editor which is nice.

bojan.matic07:04:23

the one thing I hated about spacemacs was that I wasn’t able to figure out how to increase the font size

bojan.matic07:04:28

can anyone chime in on that one?

plexus07:04:51

don't know about Spacemacs specifically but if you can eval elisp you should be able to do something like (set-frame-font "Inconsolata-17")

plexus07:04:46

if you want to change the test size just for a single buffer try going to emacs mode and use C-x C-- / C-x C-+ (control-x control-minus / control-x control-plus)

bojan.matic09:04:15

@plexus: i used emacs for a couple of years, so i am familiar with it

bojan.matic09:04:31

however evaluating the elisp code that increases font size in my config has 0 effect in spacemacs

bojan.matic09:04:08

or i think the problem was that when i put it in spacemacs init it didn’t actually change font size on startup

plexus09:04:33

I see, that's too bad. Maybe you can ask in #C09C8GRLY or #C099W16KZ

plexus09:04:41

(sorry only just catching up on the rest of the conversation, I guess it's not an acute problem for you now)

jswart15:04:11

@bojan.matic: SPC z x then use + or -

henripal16:04:32

hey all, I started learning Clojure mostly for fun. For work I use 1) Python for machine learning 2) R for stats and scientific plotting 3) C++/CUDA for computationally intensive simulations 4) still haven’t found out a way I like to deal with 100GB+ datasets. I’ve been looking to factor Clojure into my workflow, and so far I’m thinking for 1 that I could use Anglican + Clojure and meddle with probabilistic programming and for 2) that I could try Incanter out. Am I missing anything? Where else could I try Clojure out? Thanks!

surreal.analysis17:04:19

@henripal: You might be interested in Gorilla REPL for 2 (Notebook style programming, like jupyter or Mathematica) https://github.com/JonyEpsilon/gorilla-repl. For large datasets you might want to check out Sparkling, the Clojure Spark wrapper - https://github.com/gorillalabs/sparkling, or Onyx, a Clojure-native competitor - https://github.com/onyx-platform/onyx. In my experience most Clojure machine learning is done by wrapping existing Java libraries, not sure if there are a lot of native libraries, but that’s far less my area of expertise.

lich17:04:30

Personally I'm going to use deeplearning4j

henripal18:04:12

thanks @surreal.analysis ! I’ll definitely check those out. @lich, from what I understand, deeplearning4j is still cpu-only, whic might be a big handicap: http://nd4j.org/gpu_native_backends.html ?

lich18:04:38

Yes probably. But I'm just starting and see that deeplearning4j provides a framework for constructing ANNs. I don't really know maybe I can do it without a framework