There's a docstring for require.

including that excerpt from (doc require) here in case others are curious:

> (doc require)
...
  Recognized options:
  :as takes a symbol as its argument and makes that symbol an alias to the
    lib's namespace in the current namespace.
  :as-alias takes a symbol as its argument and aliases like :as, however
    the lib will not be loaded. If the lib has not been loaded, a new
    empty namespace will be created (as with create-ns).
  :refer takes a list of symbols to refer from the namespace or the :all
    keyword to bring in all public vars.

and here's a small example.

$ clj
Clojure 1.11.1
user=> (require '[nextjournal.clerk :as-alias clerk])
nil
user=> clerk/serve!
Syntax error compiling at (REPL:0:0).
No such var: clerk/serve!
user=> ::clerk
:user/clerk
user=> ::clerk/thing
:nextjournal.clerk/thing
user=> `clerk/thing
nextjournal.clerk/thing

I added a doc issue to expand coverage on clojure site https://github.com/clojure/clojure-site/issues/632

not sure "large" (size) is relevant

granularity or "complexity of update" is probably more relevant

Ok, good. I was hoping that was the case.

atoms protect updates to state so there is a tradeoff in granularity / amount of state vs contention to update - that's usually the important thing to consider

The updates should all be happening within my Clojure2d drawing thread. I don't expect contention there.

@U02UHTG2YH5 what are you putting into your atom that you thought perhaps size would be an issue?

All the state needed to draw the frames.

atoms store only 1 thing, that's the size 😂

lol

The largest is probably a set of lines to draw, usually 20ish lines. And then 30 of those sets.

I would start with one thing until that's not sufficient :)

Another point. Size kinda doesn’t matter. But one thing that tracks with size is the number of code paths that try to update the atom because it holds all the data. If 15 threads all hit the atom and change independent parts, you can get into high contention. But by itself the size of the data in the atom is irrelevant.

Thanks, everyone

and the two main techniques for dealing with it once it's not sufficient are changing atom of map --> map of atoms (for uncoordinated state) or map of refs (for coordinated state)

It has an issue tracker :) I will push the necessary update tonight

Was sick in bed for a few days 😞

oh, sorry I didn't know that

best health recovering. Thank you

https://github.com/rm-hull/nvd-clojure/pull/155#issuecomment-1378042021

thank you, I'm going to try now

🤞

(let [r #(rand-nth [1 2 3 4 5 6])]
  {(eval (r)) 1 (r) 2})

I guess it's just better to use hash-map in that case

(hash-map (random-uuid) 1 (random-uuid) 2)