So, for example, Python is pre-installed on Linux and OSX, and is a dependency of many packages in installers like apt, yum, and homebrew. It's the heterogeneity of use-cases that leads to a heterogeneity of installation approaches. R simply has never been that broadly used.
-
-
But compare Python to ruby - ruby has a single packaging system used by the vast majority of people
2 replies 1 retweet 1 like -
Replying to @hadleywickham @noamross
Sure, but e.g. Ruby also doesn't have much of a data science community... at least one of the packaging alternatives in Python (conda) grew out of unique needs there. Diverse sub-communities breed diverse tooling.
1 reply 0 retweets 4 likes -
Why does data science need special tooling? Because more data scientists use windows?
1 reply 1 retweet 2 likes -
Replying to @hadleywickham @noamross
Good question: I wrote a post that touched on that history a while back: http://jakevdp.github.io/blog/2016/08/25/conda-myths-and-misconceptions/ … Briefly: prior to ~2012, the core Python community didn't think handling compiled extensions was a priority, and compiled extensions are essential to scientific computing.
1 reply 0 retweets 8 likes -
Yeah, I’m just not convinced by your thesis that the complexity of python’s packaging ecosystem is fundamental
3 replies 1 retweet 6 likes -
Replying to @hadleywickham @noamross
My thesis is that diversity of subcommunities leads to diversity of tooling. I think the current state of affairs in Python, R, Ruby, Javascript, MatLab, etc. supports that idea.
2 replies 0 retweets 3 likes -
I’m not convinced that Python is “more diverse” than every other language
2 replies 1 retweet 5 likes -
The JVM is an interesting example of an extremely diverse platform with many communities and even different languages (java, scala, clojure) that nonetheless unified around a single packaging system (maven).
1 reply 0 retweets 4 likes -
Replying to @avibryant @hadleywickham and
Gradle, Ant, Ivy, sbt, mill, boot, leiningen?
1 reply 0 retweets 2 likes
Lots of build tools, each popular in a different community, but all use the same package format and repositories.
-
-
Replying to @avibryant @fjsosah and
And how do those handle native (C/C++/FORTRAN) libraries?
0 replies 0 retweets 1 likeThanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.