Fork me on GitHub
#datascript
<
2017-02-07
>
thedavidmeister14:02:40

@conaw what are you putting in the value of each datom?

conaw14:02:35

let’s say at least 1/2 of them are db/refs to other datoms, and half are strings that are about a paragraph in length

thedavidmeister14:02:25

well then yeah, i'd say your talking many millions for 100MB of memory

conaw14:02:39

interesting

thedavidmeister14:02:43

you could test it pretty easy

conaw14:02:43

@tonsky do you think a datascript app would perform comfortably well with many millions of datoms? Has anyone seen this in the wild?

conaw14:02:50

yeah with generators

thedavidmeister14:02:58

just transact a bunch of data into a db

conaw14:02:59

good point

thedavidmeister14:02:43

from my experience (with far less than millions of datoms) you'll get different perf characteristics with the different APIs exposed by datascript

thedavidmeister14:02:59

e.g. queries vs pull vs datoms

thedavidmeister14:02:30

but i don't know how they scale differently

thedavidmeister14:02:49

i'd assume, for instance, that the query building part of running a query doesn't change with additional datoms

conaw14:02:09

interesting

thedavidmeister14:02:26

just try it out 🙂

conaw14:02:57

yeah I’ll report back, 🙂

thedavidmeister14:02:11

pragmatically speaking though, you're probably not going to be rendering and updating millions of DOM elements

conaw14:02:48

nope, this is just a thought experiment on how large an app could get with #nobackend