This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2017-08-19
Channels
- # aws (9)
- # bangalore-clj (1)
- # beginners (6)
- # cider (17)
- # cljs-dev (9)
- # cljsrn (31)
- # clojure (70)
- # clojure-spec (19)
- # clojurescript (122)
- # cursive (1)
- # datomic (20)
- # dirac (8)
- # fulcro (19)
- # hoplon (25)
- # instaparse (2)
- # lambdaisland (3)
- # luminus (8)
- # lumo (10)
- # om (22)
- # parinfer (25)
- # protorepl (2)
- # re-frame (35)
- # reagent (1)
- # test-check (5)
- # unrepl (6)
I was once fascinated with Clips - a expert system tool based on the Rete Algorithm that has appeared in Clojure as Clara https://github.com/cerner/clara-rules and now I am seeing it being used for UI: https://github.com/CoNarrative/precept and https://github.com/arachne-framework/factui this may be somewhat tied to react but I thought it would interesting to explore in Hoplon — by someone smarter than me of course
you could use something like Factui with Hoplon by wrapping the queries into cells. Have one input cell with the fact db. defrule
then become analogous to a formula cells.
yes i'm quite interested in precept
i've been using datascript in cells so far
but i understand precept could be much more efficient than that right?
yeah, the rete engine trades off memory for speed - it will choose the rules to evaluate efficiently based on the facts that changed
i think that could be very beneficial based on my experience so far
datascript in a UI tends to lend itself to many small changes because of the datoms structure
and then you're re-running all these queries, maybe even queries for other routes if they're cached in a tpl
but i don't think that it actually takes up that much memory out of the box, it's just churning through queries that don't change
so the tradeoff might be quite smart
you'd need many datoms to use up even 100mb of RAM i think
sure, for a single client, unless it’s some sort of a physical simulation, the amount of data shouldn’t be that high
yeah i'm thinking "typical" web dev
yeah that's what i haven't done yet
figured out how to hook precept into cells
i imagine it would almost be a 1:1 drop-in replacement at that point
ymmv but so far i've found datascript to be a performance bottleneck much more often than hoplon
and the workarounds usually involve dropping down into raw datoms manipulation
so from that perspective precept/javelin integration is an exciting future project 🙂
⚠️`[hoplon/brew "0.3.0"] deployed to clojars!` ⚠️ This repo has been stable for a while.