#openspending

last thing I've thought about is authentication and I've basically settled on OAuth

pudo forked an example oauth2 provider and rebranded it as citizenID: https://github.com/pudo/citizenid

I like the idea behind it, civic apps with the same login

and it fits perfectly into the micro-service scheme

so we would just use the access tokens from citizenID for authentication in our microservices

we could even make the experience as seamless as possible by allowing Resource Owner Password Credentials Grant: http://tools.ietf.org/html/rfc6749#section-4.3

so we would just have the "normal" username+password fields

and citizenID would treat "OpenSpending" as a confidential and trusted app

it could also be implemented with a button or something (I'm thinking the web interface)

which would be what is usuallly done e.g. with the "login using github" or whatever is out there

this would then also be usable by the api

hi nathanhilbert

Hey sorry to be late to the party. Just wanted to check if anyone was still discussing

nathanhilbert: no problem, I just finished my discussion notes a minute before you popped int

I'll check the archives

nathanhilbert: starts about here: https://botbot.me/freenode/openspending/2015-02...

pwalsh: did you have any comments on the OAuth thing?

about the oauth. In general, I'm all for it. I didn't look too deeply into citizen ID, but I've done this stuff a few times by now. Would I would like to see is citizen ID (or whatever) not only be a provider, but a proxy for other providers. what that means, is, like on Stack Exchange, you can create a stack exchange account "natively", or you can create one with yer google/twitter/fb/what-have-you

*What I would...

I thought, in general, that this was a good idea for OK services, but I like pudo's vision better being that it is for a range of "citizen services" and may see adoption with other NGOs etc

pwalsh: yeah I haven't looked at lastuser which you pointed me to

but that sounds like the proxy you're talking about

lastuser might be enough, citizenID is not developed enough

Nigel from OK worked for the company that builds lastuser

FYI

hah, I did not know that

nigelb: is that true?

I only heard about it through him

would you recommend lastuser?

hey Stiivi

is there anything else we want to discuss

I have to run out in a few minutes

but if there is nothing else we can conclude the meeting here

(not that this should stop discussions in the channel)

there will be another one tonight at 23 GMT

back now

this meeting still alive?

I'm running out after I finish my coffee

pudo: will you be around at the "US edition" meeting this evening?

yes

awesome

sorry to miss this, very interesting scrollback

I don't have working code like you, but I've thought a bit about things ;)

alright, I'm off for now

feel free to go on without me :)

pwalsh: hey

:)

you're doing the uk25k stuff, no?

trickvi: It is indeed true :)

hey pudo

yes indeed

how's it going?

trickvi: I have to look at the codebase again to see if I would recommend it. I don't remember much of it.

that should give you a better idea of shitty ETL than you ever wanted :)

I'm surprised to read that trickvi wants developers to run the whole microservice thingy locally

I really can't see it's benefit at OpenSpending's scale

we're gaining complexity for no particular purpose

pudo: that's not what I said I wanted, but that's something we will need to think about, I'm hoping mocked interfaces is good enough for most devs, but for some devs we might need to offer the "whole package" (those interested in the whole system)

anyways, we'll be starting in about 13 minutes :)

alright, hello

good evening, good morning or good day, have your pick of a time zone aware greeting

is anyone present for this second dev meeting today?

let's give possible latecomers a few more minutes

I guess nobody is coming to this meeting