This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2015-08-19
Channels
- # admin-announcements (14)
- # architecture (1)
- # beginners (21)
- # boot (301)
- # clojure (93)
- # clojure-brasil (56)
- # clojure-india (4)
- # clojure-italy (11)
- # clojure-japan (25)
- # clojure-russia (16)
- # clojure-sg (1)
- # clojure-uk (5)
- # clojurescript (64)
- # clojurex (4)
- # clojutre (7)
- # cursive (18)
- # datascript (35)
- # datomic (212)
- # editors (27)
- # emacs (38)
- # events (9)
- # funcool (21)
- # hoplon (105)
- # jobs (2)
- # ldnclj (41)
- # ldnproclodo (4)
- # liberator (89)
- # luminus (5)
- # off-topic (32)
- # om (39)
- # onyx (1)
- # re-frame (9)
- # reactive (1)
- # reagent (34)
- # testing (42)
- # yada (4)
I lose at almost everything (lookups, insertion, structural sharing percentage) but win a lot on scans
here’s a perf comparison between 0.11.4 and 0.11.5, when I upgraded JVM implementation from sorted-set to BTset
@tonsky: ok, so it’s for performance reasons. Have you tried to compare perf between clojure’s red-black trees and data.avl?
Logically, representing ‘datoms’ as a sorted set of facts (with different sort orders, such as :avet :eavt etc) is the correct abstraction for implementing datalog? Or is there something missing from the ‘sorted set’ abstraction?
@tonsky one more question (sorry to bother you). Is the benchmark code available somewhere?