Fork me on GitHub
#cljs-edn
<
2016-05-01
>
shaunlebron19:05:22

i feel like my visibility of the problem-space is getting better, not convinced of any solution yet

shaunlebron19:05:53

just a proposal, let me know how that jives with your direction with planck

shaunlebron19:05:10

i still have to think about boot, the last example left

mfikes20:05:21

@shaunlebron: Yep. Looking good. (As an aside, your mention of npm reminded me of an idea that Planck could also conceivably support :module-type :commonjs by shelling out to Closure. In which case, Planck would continue to just follow the existing config standards and not innovate there.)

shaunlebron20:05:53

@mfikes: if you could shell out to closure for that, would be interesting to see if it could become a full build system, haha

mfikes21:05:15

Yes. I was pondering that the other day. Planck already uses itself to AOT its own macros. http://blog.fikesfarm.com/posts/2016-02-03-planck-macros-aot.html It is 1 step away from being a build system. (But, there’s probably a lot more to a full-fledged build system.) You could almost get philosophical about the intermixing of runtime / compile time / read time (Paul Graham’s item 9). In short, since Planck doesn’t need to AOT for deployment to the web, it currently enjoys freely blurring the distinction, even having eval http://blog.fikesfarm.com/posts/2016-01-22-clojurescript-eval.html. Part of my philosophical view is that ClojureScript was heavily affected and shaped by its need to AOT for a remote JS environment, one that I’ve never targeted. (I’ve only used ClojureScript outside of the web.)

mfikes21:05:09

In short, bootstrap feels a lot more like what I think of when I think of LISP simple_smile

shaunlebron21:05:00

that’s really interesting

shaunlebron21:05:38

glad that planck is there to see that distinction

shaunlebron21:05:30

@mfikes: just re-read those articles a few times. very cool stuff. i feel like i understand Lisp a little better now

shaunlebron21:05:00

i kinda missed all that stuff because my experience with Lisp has been in this constrained browser environment

shaunlebron21:05:05

might be a dumb question, but is clojure self-hosted?

shaunlebron21:05:12

because it has eval?

mfikes21:05:28

Yeah, I’ve been meaning to write a “philosophical” post discussing the two extremes that ClojureScript now affords us. Will scratch that together.

mfikes21:05:01

I think to be self-hosted, you need to be able to compile yourself. But but that’s an interesting question, given “hosted” languages like Clojure and ClojureScript.

shaunlebron21:05:06

i get lost in the layers of it all when thinking about it