Stream: t-compiler/wg-nll

Topic: quantify-stackoverflow


nikomatsakis (May 18 2018 at 17:28, on Zulip):

Hmm @Jake Goulding your recent question got me thinking. Do you the maintainer folk of stack overflow make any effort to quantify the sorts of questions you get?

nikomatsakis (May 18 2018 at 17:28, on Zulip):

it seems like that could be awesome data about the kinds of patterns that trip people up

nikomatsakis (May 18 2018 at 17:28, on Zulip):

features that are confusing

nikomatsakis (May 18 2018 at 17:28, on Zulip):

etc :)

Jake Goulding (May 18 2018 at 17:30, on Zulip):

There's loose quantification via the tagging system. If you search for [rust] you see a list of related tags, for example:

lifetime × 515
traits × 408
rust-cargo × 394
borrow-checker × 301
generics × 251
iterator × 225
macros × 197
string × 196
ffi × 191
vector × 162

nikomatsakis (May 18 2018 at 17:30, on Zulip):

interesting.

nikomatsakis (May 18 2018 at 17:30, on Zulip):

I should think a bit more carefully about the kind of queries I'd like to run

nikomatsakis (May 18 2018 at 17:30, on Zulip):

/me senses a project he could use to learn him some machine learning

nikomatsakis (May 18 2018 at 17:31, on Zulip):

though surely somebody is doing this already

Jake Goulding (May 18 2018 at 17:32, on Zulip):

I try to ensure links to issues are present when there's an actual bug, but it doesn't sounds like you want bugs per se, but more of conceptual roadblocks

Jake Goulding (May 18 2018 at 17:32, on Zulip):

There's also http://data.stackexchange.com/ if your SQL skillz are good. It's a weekly dump of the entire database.

nikomatsakis (May 18 2018 at 17:33, on Zulip):

I mean... I know as much SQL as the next person, I guess :)

nikomatsakis (May 18 2018 at 17:33, on Zulip):

something something SELECT something something :)

nikomatsakis (May 18 2018 at 17:33, on Zulip):

thanks

Last update: Nov 21 2019 at 14:00UTC