News for hackers. Stuff that matters.
Re:Why I Live in IRC (2015) by catdog on Thursday August 11, 2016 @09:34AM (#12268063)
IRC originates from a time where nobody was thinking about multiple devices, or e.g. not encoding in ASCII. Speaking of nonstandard: There seems to be no canonical specification of IRC itself. Nevertheless it's really not that bad in practice if you use a bouncer or something similar like Quassel.

> There's no reason the platform has to be closed - I'm friends with the people working on http://matrix.org/ (though not involved in it myself) and hope they succeed.

There is also XMPP (and has been for years) which does not suffer from many of IRCs shortcomings.

Re:Why I Live in IRC (2015) by stonogo on Thursday August 11, 2016 @11:25AM (#12268973)
> There seems to be no canonical specification of IRC itself.

Is this a joke? Did you think all this stuff just accidentally worked? I'm not sure how seriously to take your entire message because of this bizarre idea.

https://tools.ietf.org/html/rfc1459

https://tools.ietf.org/html/rfc2810

https://tools.ietf.org/html/rfc2811

https://tools.ietf.org/html/rfc2812

https://tools.ietf.org/html/rfc2813

Re:Why I Live in IRC (2015) by catdog on Thursday August 11, 2016 @01:42PM (#12270244)
Sure it's standardized enough to generally work but to my knowledge only RFC1459 is widely seen as authoritative. The protocol evolved quite a bit over the years but a lot of the extensions covered (and maybe not covered) by those later RFCs are handled to some extent differently between IRCd implementations. There at least seems to be some effort [1] now to fix some of IRCs issues and bring the whole mess more in line again. As IRC is still widely used and probably there to stay I hope it will work out.

[1] http://ircv3.net

Re:Why I Live in IRC (2015) by zeveb on Thursday August 11, 2016 @10:19AM (#12268401)
> There is also XMPP (and has been for years) which does not suffer from many of IRCs shortcomings.

In return, it suffers from its own shortcomings, mainly XML and second-system effect (leading to too much complexity & too many options).

What we really need, I think, is exactly what Slack, Mattermost &c. give: persistent, multi-device, multi-group chat. I think that solving that problem in a nice, small, secure protocol (one that doesn't imply a browser-first way of doing things — and indeed, 'secure' probably prevents using a browser in the first place) is more than enough of a problem for someone to solve.

Heck, even just the multi-device shared-identity problem is surprisingly difficult.

Re:Why I Live in IRC (2015) by lmm on Thursday August 11, 2016 @09:41AM (#12268104)
> not encoding in ASCII

IIRC the original standard specifies a particular 8-bit character set (not ASCII), though the broader point stands.

> There is also XMPP (and has been for years) which does not suffer from many of IRCs shortcomings.

XMPP is useless for multiple devices: there's no way to receive messages on both (some servers will deliberately violate the spec here because the spec is stupid), and the federated history spec has been under development for about a decade now. Groupchats don't work properly with federation and have terrible UX in every client I've tried. Everyone assumes XMPP must be good because it's open-source (which is usually a pretty good heuristic) but I've come to the conclusion it's actually a really bad standard.

Re:Why I Live in IRC (2015) by catdog on Thursday August 11, 2016 @10:47AM (#12268628)
> XMPP is useless for multiple devices: there's no way to receive messages on both (some servers will deliberately violate the spec here because the spec is stupid)

XEP-0280: Message Carbons and there was that priority thing before. It somehow worked in the past and now it's definitely solved.

> Groupchats don't work properly with federation

The only annoying problem i can think of is when the server hosting the MUC restarts clients seemingly don't notice that they have to rejoin.

> and have terrible UX in every client I've tried.

Maybe you tried terrible clients? Unfortunately are few now which aren't terribly behind regarding protocol support.

> and the federated history spec has been under development for about a decade now.

I think thats the real problem. "Nobody" implements newer XEPs, especially if they are not finalized yet but how should they finalize if nobody puts them to trial? Everybody seems to have settled in screaming out what does not work so well but instead of helping to fix it they halfway reinvent the wheel or switch to centralized proprietary solutions which may have nicer looking clients but often don't even try to solve harder problems like multi client support. I didn't have a look at Matrix yet, maybe it's worth it but there is always the danger of creating yet another standard [1] (which may not even be as perfect as originally thought in the end either).

> but I've come to the conclusion it's actually a really bad standard. But is it really that bad (especially at the basic fundaments) to justify something completely new? XMPP at least provides a lot of ability to evolve without breaking everything that was.

[1] https://xkcd.com/927/

Re:Why I Live in IRC (2015) by Arathorn on Thursday August 11, 2016 @05:32PM (#12271711)
hopefully Matrix avoids falling into the cliché of xkcd 927 by focusing on defragmenting the existing silos - building a metanetwork between IRC, Slack, XMPP, etc etc. So rather than saying "hey everyone, throw away your current stuff", we say "hey everyone, build a bridge into Matrix - it's fun and easy (100 LOC for the slack bridge: https://github.com/matrix-org/matrix-appservice-bridge/blob/....

If you want to use Matrix natively via weechat or whatever then so much the better, but the intention is to be useful as way more than just another msging protocol. It's meant to more be a big open decentralised database for conversations and other realtime data - more in common with IPFS or even the Web itself.

[disclaimer: I work on Matrix]

[p.s: can someone work out how to send Matrix $5 every time someone quotes xkcd 927 at us...?]

Re:Why I Live in IRC (2015) by corobo on Friday August 12, 2016 @05:44AM (#12274319)
> p.s: can someone work out how to send Matrix $5 every time someone quotes xkcd 927 at us...?

I'm sure when Matrix is the go-to chat protocol you'll get more than that. Until then it is sorta fitting..

Re:Why I Live in IRC (2015) by Arathorn on Friday August 12, 2016 @10:16AM (#12275798)
sigh no, the point is that it doesn't fit because Matrix does not try to be "yet another chat protocol that fits everyone's use cases". It aspires to be a bridging protocol that links other chat protocols together rather than replacing them.
Re:Why I Live in IRC (2015) by corobo on Friday August 12, 2016 @10:26AM (#12275897)
I'll be honest I've not scoured your entire site but I'm just not seeing that description (and didn't at all earlier, I've since looked further).

From what I gathered you want to be the email of chat protocols. The only mention of wanting to bridge other protocols is a quick mention half way down your FAQ page.

Re:Why I Live in IRC (2015) by Arathorn on Friday August 12, 2016 @05:26PM (#12278922)
Guess this reflects that the FAQ is stale and needs some love. The second item there is:

> What is Matrix’s Mission? Matrix’s initial goal is to fix the problem of fragmented IP communications: letting users message and call each other without having to care what app the other user is on - making it as easy as sending an email.

The whole point is that the defragmentation is done by bridges. The analogy to email is not in terms of having One True Protocol that everyone uses like SMTP, but that it's possible for a user on app A to communicate with another user without caring what app they use.

Thanks for highlighting that we need to fix the FAQ!

Re:Why I Live in IRC (2015) by corobo on Friday August 12, 2016 @06:42PM (#12279303)
Fair point, I think I read "without having to care what app" as meaning a Matrix-based app, whether official or not.