Conversation

100% agree: empirical guidance language design most needs isn’t “which keywords are easiest to parse”, it’s “how do we expand the terrain in which the user has enough situational comprehension to modify other people’s code and/or respond appropriately to compiler diagnostics” twitter.com/chrisamaphone/

This Tweet is unavailable.

Graydon Hoare

@graydon_pub

Sep 6, 2018

(At least, outside of teaching/learning languages. Every bit of friction surely matters there! But I'm one of those weirdos who is fine with, even thinks it's probably good to have, separate languages for teaching/learning and production use. But that's a different ball of wax.)

Brendan Zabarauskas

@brendanzab

Replying to

@graydon_pub

I love teaching langs, but I'd like to see some research into how students feel about learning them. Just from what I've seen, maintaining adoption of them (in courses) seems like it is hampered by the students own worries.

5:34 PM · Sep 6, 2018Twitter Web Client

Replying to

and

Been thinking having 'teaching modes' for languages, or 'beginner preludes' might be potential solutions. I dunno.

Replying to

At least one thing seems to work to extend/expand adoption of a teaching lang: make them passable as the upper-half of Ousterhout's dichotomy. Casual/forgiving/scripting/small-program level. Audience is quite big for that! Probably majority of programs by volume. See VB, Python.