Agreed... now I have 2 concrete concerns: 1. Where will the time come from to teach coding? (Not sure that "From the time not spent learning SPSS clicking" will be sufficient.) It's easy to have ideas for adding to a curriculum, but nobody wants their stuff to have less time.
-
-
Replying to @sTeamTraen @o_guest and
ex1 <- read.csv("ex1.csv", header = T) plot(ex1$iv, ex2$dv) lm(dv ~ iv, data = ex1) Is the learning curve at that level really do high? (Genuine Q - I use it every day so can't judge!)
2 replies 0 retweets 2 likes -
Replying to @joejps84 @sTeamTraen and
Yes, there's a lot to unpack there. You first need to teach them what a variable is. You also need to teach them what a function is, and that code executes in order. I think a lot of people forget how hard it is to code if you have literally never done it before.
1 reply 0 retweets 10 likes -
Replying to @bradpwyble @joejps84 and
In the 1980s I taught a course for engineers using our CAD product which had a powerful programming language built into it. They understood sin(x) no problem, but the idea of writing your own fn(x, y, z) that returned a number was often really hard for them to understand.
1 reply 1 retweet 3 likes -
Replying to @sTeamTraen @bradpwyble and
Other issues: - Local variable scope (different in every language!) - Strings versus character vectors versus names - Vectors versus lists Many of these things require a "mathematical/logical" mind, which may be partly orthogonal to talent for statistics.
1 reply 0 retweets 4 likes -
Replying to @sTeamTraen @bradpwyble and
Example of the scope problem. > f<-function() {print(X); X<-X+1; print(X)} > f() Error in print(X) : object 'X' not found > X<-3 > f() [1] 3 [1] 4 > print(X) [1] 3 Now replace X<-X+1 with X<<-X+1. Think how much understanding you need to master this.
2 replies 0 retweets 1 like -
Replying to @sTeamTraen @bradpwyble and
I've programmed in ~25 languages, ~15 of those for pay, and I found some R stuff a bit tricky at first. For example, imagining how the parser must work for constructs like formulas to be possible (as opposed to as.formula(string)) is a little mind-bending.
2 replies 0 retweets 1 like -
Replying to @sTeamTraen @bradpwyble and
a context sensitive parser must be a nightmare to debug
1 reply 0 retweets 1 like -
Replying to @sharoz @sTeamTraen and
When I taught a small class of undergraduates to do neural networks from scratch (no library, just for loops) we didn't do scoping and they coped.
1 reply 1 retweet 7 likes -
From a pedagogical perspective it's not ideal to introduce scoping straight away anyway.
2 replies 1 retweet 6 likes
Also in high school & in computer science A level, I recall scoping was not introduced until much later. This is a common pedagogical principle. You teach, for example, about 3 states of matter (solid, liquid, gas). You don't go "oh, yeah, there's also Bose-Einstein condensate".
-
-
I have the slides for the undergrads, if anybody is curious. They had zero programming experience in Python and only 2 of 15 had coded before.https://github.com/oliviaguest/connectionism …
1 reply 1 retweet 10 likes -
They all got a first or a two one, and not because I'm too nice.
0 replies 0 retweets 4 likes
End of conversation
New conversation -
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.