#dat

ogd: almost cut the dat minimal example in half with a default swarm module: https://gist.github.com/joehand/1e644adc3cc43ae...

the default swarm module is pretty basic though. not sure if that will be worth having unless you expect having those defaults in a lot of places, then it'll be easier to change in one place. https://github.com/joehand/datland-swarm

jhand: nothing wrong with a small module! 😁

jhand: wow

jhand: heres a possible alternate way to do it

jhand: however mafintosh and i think your way is definitely simpler. only issue is if we change discovery-channel/discovery-swarm often it might be annoying to have to bump datland-swarm all the time

jhand: i guess a third way would be to just have it return a json object with our config

ogd: ya that makes sense. Didn't think about that dependency issue. Walking home, I'll look at that code there

jhand: i think yours is the simplest which is nice

jhand: heres a small tweak:

ogd: ya that one makes sense.

jhand: you think its still easy/intuitive enough?

jhand: in general though its super awesome all the code fits on one screen now

ogd: I think so. The other option would be requiring it directly in discovery-swarm. but then you are forcing people a bit more to use defaults.

ya!

ogd: would that make sense to pass drive as the first argument?

jhand: thought about that too but seemed kinda weird

ogd: ya definitely

jhand: cause it makes it seem like its doing more than just instantiating

jhand: like 'oh im giving it my swarm it must be doing something complicated'

ogd: right. it seems like its hiding stuff, but its just hiding 1 line...

thats what dat links are o/

TIL there was a name for security by obscurity URLs

first very rough draft of the mutable (or appendable) feeds for hypercore, https://gist.github.com/mafintosh/d45d847a70284...

What's drive.core.id?

I was just wondering the same thing =)

its the peer id

it should be exposed on the drive itself

i'm actually not sure if we need it in the swarm anywore

do we need it in that example? it looks like discovery swarm will set a random ID if we don't pass it one

jhand: i can refactor that away yea

jhand: the swarm uses it to detect if it's connecting to itself

which is a surprisingly tricky problem

mafintosh: ah okay.

jhand: but it is exposed on the peer stream so the drive could just read it there

jhand: if you wanna pr that i'll merge it!

:)

mafintosh: sure I'll take a look

shama: cool i'm playing around with ids

shama: it definitely changes the behavior, though i'm still wrapping my head around it

shama: and just saw the style addition, works great!

(Blahah) @ogd @mafintosh @karissa can you envisage the entire scientific publishing infrastructure resting on dat?

ya

except search

(Blahah) indeed

but file hosting yes

(Blahah) let's plan what that looks like

(Blahah) there are lots of decent components out there, for authoring, reading, peer review

(Blahah) but they would need to work on the dat storage layer

@Blahah i think it'd be nice to have a ui for PRs once we have dynamic feeds

ogd karissa mafintosh: I've created a wiki that walks through an architecture for composing elements: https://github.com/shama/bel/wiki (Next up is one outlining some ideas for handling CSS)

btw, bummed that donut.js is happening on the day I travel for YNAB's retreat :(

shama: ah dang

shama: awesome!!

shama: yeah great overview!

jhand: me and mafintosh are hacking on division st

ogd: cool i can head there in a bit. where you at?

jhand: dapper & wise