Fork me on GitHub
#data-science
<
2017-10-30
>
didiercrunch01:10:59

lein-jupyter now includes code highlighting and forward command line arguments to jupyter notebook (useful if you need to start jupyter on a different port then the default one). On my computer parinfer works but I did not have time to create the install script. You will need to re install the kernel to use the code highlighting.

sparkofreason01:10:04

Awesome! Can't wait to try it.

whilo16:10:00

@michaellindon here i am 🙂

michaellindon16:10:57

hey hey welcome to the party xD

michaellindon16:10:14

a lot of good discussions here, no history tho :<

michaellindon16:10:37

people here should also be aware of this other message board https://gitter.im/metasoarous/clojure-datascience

whilo16:10:12

yeah, slack without history is a bit ...

michaellindon16:10:34

maybe everyone can migrate over to gitter

michaellindon16:10:38

is history on gitter better?

whilo16:10:33

yes. also gitter is open-source, which is a big thing for infrastructure code in my opinion.

whilo16:10:27

@didiercrunch nice work! do you have some comparison to gorilla? i think jupyter has more features, but is far less hackable. the latter is barely relevant in most cases, but i appreciated that it was written with a clojure stack.

michaellindon16:10:48

Perhaps its good if people are aware of the presence of https://gitter.im/metasoarous/clojure-datascience in addition to the slack channel

michaellindon16:10:24

i think there are more people on the slack channel, but there are also some good discussions going on in the gitter channel.

whilo16:10:00

i think christopher has started this channel once a long-time ago and a few people joined over time. the slack channel is more official in a sense

michaellindon16:10:10

I was talking about this with someone a while back... there argument was that the clojure community as a whole was on slack, so it would be good to keep the data-science community there also

michaellindon16:10:43

theres also meant to be a google groups lurking around somewhere

michaellindon17:10:13

is no history such a bad thing? Most of the time when I search for something, theres a post from say 2015, and i dont know if the answer is even relevant for the current day

whilo17:10:41

hehe, it is not a bug (restriction), it is a feature

whilo17:10:26

well, i prefer to decide for myself. one big advantage over IRC of the webchats is the history

whilo17:10:54

this is just an arbitrary restriction of slack to make money

whilo17:10:14

which i understand, but it is a problem for open-source projects

didiercrunch17:10:18

Can't we have a bot that record the history?

didiercrunch17:10:54

We could serve the history as static files or even from github

didiercrunch17:10:22

Cool! So we have history

michaellindon17:10:25

ok so this history is nicely covered

didiercrunch17:10:45

Peraonally, I like being part of the general clojure community

Quest20:10:27

I'm a lurker here, as I don't do data science, but intend to at some point in my career. I appreciate the chance to read through the channel musings, though I doubt I'd follow it if the channel separated from Clojurians. (Not saying whether the channel should move or not. The use of the channel's community matters more. But perhaps lurkers like myself should be a consideration... thinking back to the machine-learning call to arms that was Conj's closing speech.)

michaellindon17:10:12

come join us @whilo 😎

chrjs17:10:04

One of us… one of us… 👥 👥 👥 👥 👥 👥

whilo18:10:20

Hehe 😎

whilo18:10:31

Assimilated, resistence is futile.

michaellindon18:10:32

its great to have you on board, I think you could really contribute here

didiercrunch23:10:56

@whilo gorilla is very good but afaiac jupyter is much more mature. The issue with jupyter is not its customizability but with the distribution. You can customize whatever you want in jupyter and it is fairly easy; simple zero-mq based server side and classical requirejs/jquery client side. The issue is to distribute extensions. I do not know what [jupyterlabs](https://github.com/jupyterlab/jupyterlab) does to address the issue.