XSF Discussion - 2014-02-25

Mozilla also accepted and have some XMPP-related tasks for Instantbird

Santiago26: that's excellent news. Do you have a link?

https://wiki.mozilla.org/Community:SummerOfCode14#Instantbird

looks like we need a compliance testing suite real-soon-now

simon: ah, i spoke to florian who reported the webrtc stuff at fosdem :-)

'reported'?

he's listed as reporter

probably proposed those items

The 'additional protocols' one is funny

should there be an xsf blog post how happy the XSF is to see all those projects? :-)

fippo, i'm +1 for that...while it's unfortunate the XSF itself doesn't mentor any of the projects it's nice to see adoption of XMPP style projects in other mentoring orgs

tobias: that would be the message ;-)

i'll see if i can write something up... after another round of skiing

I also think we should try to support those efforts

fippo: do coordinate with Laura, by the way

fippo, you make it sound like you're on holiday tztztzt

I guess "break a leg" isn't appropriate now?

hehe

will do, i typically need someone to wordsmith anyway (-:

We need a proof-of-work protocol for registering user accounts and for sending messages beyond a given limit!

Account registration combined with crypto currency ;)

Every account you register you gain bitcoin?

yeah...so each client will come with a 20 gig hash chain :)

even though my comment was tongue-in-cheek, it might work out to reduce spam. Also, it could be used by the server admin to delegate bitcoin mining to clients, repaying for the server operation.

And give server operators an incentive to actively welcome clients that are part of a botnet.

Sounds good. Who's going to write up the spec?

botnets are using tor already... why not do some mining?

We filled the … in 1. operate jabber.org, 2. enable registrations again, 3. …, 4. profit!!!

intosi: btw, that is not quite true. the botnet operator will rather be incentivized to mine their own bitcoins, instead of sponsoring the xmpp server

we had throttling in xep-0198. what about letting the server send a "now please calculate this for me, or I won't forward your last message" to the client?

I'm sure the mobile client developers will love the idea of even more power consumption.

Or the IoT-guys, for that matter.

intosi: Although if you're doing IoT you're probably not doing it on free public servers.

intosi: having your smartphone spin a wheel for 10-20s when registering is ok IMHO... and having your account throttled is not too much as well

Or maybe you are, I don't know.

Kev: that might be a fair point, although I don't know either.

I've asked in jdev, but I'll ask here - anyone got a real-world use-case for Teemu's two-factor thing? I don't know if it's just me, but it's still too opaque to me.

Ge0rG: it might not be an issue at registration. And if you normally don't hit this unless you send a large number of stanzas, it wouldn't even be that big an issue for those either.

dwd: I don't understand it.

intosi: that's what I'm saying

Ge0rG: right. And I presume you would suggest just falling back to rate limiting in the old-fashioned way if a client doesn't support this extension.

intosi: yeah.

intosi: except for registration, obviously

Obviously. No IB registration if you don't support the extension.

Somehow that feels like another step in an arms race to me.

But that might just be the nature of these problems.

intosi: I'm not quite sure if this is just another step or a different order of magnitude.

If it takes 10-20 seconds to do the work before you can slowly send 1000s of spam messages from a botnet drone, the cost might not be high enough to make it an order of magnitude. It might be enough to make botnet attacks harder.

But this isn't the biggest problem model we have at the moment.

It's signing up accounts to send a few messages each.

Kev: that's true.

intosi: I'd say that once you run into throttling, you have to work 10s per message or somesuch.

of course, 10s on a smartphone is different from 10s on a desktop

btw, is there a way to throttle if a user sends many messages to different, not-in-roster, users?

sure, all impl. specific

My wife and daughter are making fudge and chocolate welsh cakes in the kitchen.

I just heard my wife complain she's burnt one. My daughter immediately says, "Dad, you can have another!".

Well, thanks.

heh

that's a great form of rate limiting

:)

Fudge and chocolate welsh cake sounds like a wonderful thing.

<proof workload='bake_cookies'/>

I'm only just down the road, you know.

"Can you create an account for me?" "Sure, bring me a cookie.". Don't see the difference from asking for a cup of coffee 15 years ago if soneone needed something done urgently.

intosi: yeah, token based authorization is awesome

Kev, It's either/or. Theer are some fudge-chip, and some choc-chip.

Kev, The fudge-chip ones are awesome when they're still warm.

You see where they made their first mistake.

Kev, Well, the choc-chip ones have molten choc in them until they cool. It's not quite the same.

Kev, Also, you could eat both at once.

dwd: now that you have cookies, can I have yaxim-muc?

Ge0rG, I'm almost emerging from my current day-job workload.

Is the editor meeting in editor@muc.xmpp.org or council@muc.xmpp.org, email and calendar invite disagree

ah!

I am waiting in editor...

oh that's right we have editor!

Me too (currently)

let's use the editor room

Reading XEP-0001, I note that §5 says that after the submission is made, "If no member of the XMPP Council objects to publication of the proposal within fourteen (14) days or at the next meeting of the Council" it's accepted and published.

That's not quite the same thing as the Council has done recently (it's been considered for a maximum of 14 days after the next Council meeting). I suppose that's fixable de-facto within the process by a Council member objecting de-jure until the timeout, but still.

dwd: Ah, it is what Council are doing.

Council are holding two-week long meetings, starting in a MUC, and ending on teh mailing lists :)

That's also a way of expressing it...

It's what we do, though.

We have a realtime portion, and then allow voting in the meeting to continue for two weeks, to allow for absent members, or those who couldn't get stuff reviewed in advance.

And as far as I could tell, there's no problem with this.

Yes, I entirely agree.

I was aware of the XEP1 language, and didn't think there was anything conflicting about what it said and what we did.

Ah, there I don't agree entirely. At best it depends on how you interpret it. I'd rather we just documented what we do clearly.

gmgm

errrr...

fippo: how is your week going?

stpeter: code-ski-code-ski-code-sauna-code... quite productive ;-)

fippo: sounds delightful!

Sounds awesome.

Next summit in Lapland?

:)

I tried skiing once and loved it. I'm somewhat concerned that trying it again, especially given my lack of fitness, would be an invitation to destroy myself.

i'm sure there are easy tracks

I prefer nordic skiing

I don't fancy hurtling down mountains

Tobias: Sure, it's not as if I did more than blue runs last time.

stpeter: I don't think one would describe what I did as hurtling.

Right, mail sent to my registrar asking them when they will allow submission of DNSSEC DS records for the NL TLD. They support it for .net

DLV kinda sucks, and the XMPP Observatory (rightfully, I think) doesn't check it.

what about DNSSEC on .im?

dwd: I think you are getting to cought up in process

caught

Ge0rG: there are efforts underway to achieve that

hmm, I haven't sent my yearly inquiry to the nic.im folks about DNSSEC support

stpeter: I think others might have.

dwd is working on that, as far as I know.

the more the merrier

Well, not really.

maybe I need to hop on over to the Isle of Man when I visit the British Isles next week :-)

stpeter: dwd has been working on this

allrightie

http://mail.jabber.org/pipermail/operators/2013-November/001963.html

However, I don't think he has gotten a meaningful response other than an receipt acknowledgement

right

that's why I figure it can't hurt too much for them to know that there's generalized demand, not just one lone person

I believe dwd approached them on behalf of the XSF, but I might be mistaken.

this is true

that doesn't seem like the kind of thing that Dave would do

since he's Mr. Process

stpeter: it has been discussed in the Board meetings

and the Board or membership didn't officially deputize him blah blah

oh, I must have missed that meeting

I can't find a reference, though, so maybe I'm halucinating

ralphm, stpeter - we discussed. However, it probably wasn't part of a board meeting.

ralphm, stpeter - And FWIW, I'm not so much interested in slavishly following a process as ensuring the process we actually follow is documented correctly.

ralphm, stpeter - So in the case I noted, some author could insist their ProtoXEP should be published as a XEP according to XEP-0001 because it no objection has been raised for 14 days from its submission. The wording is such that an argument could be made that what we do is acceptable, but it'd be an unpleasant discussion at best.

ralphm, stpeter - I thought that given the new editorial team, I'd run a pass over XEP-0001 and see if these minor issues could be cleaned up easily with an edit or two.

ralphm, stpeter - What I'd really hate to happen is the kind of process drift in the IETF, where theory and practise often diverge quite badly.

dwd: yes, that all makes sense

Update for the goto fail-issue for OS X just out. No sign in the release notes, but 10.9.2 apparently fixes it.

apparently it fixes it. Love that the fix could potentially be MITM'd

I'd be really surprised if OS X updates don't come signed separately.

dwd: nod

xnyhps: let's chat in London about how to manage xmpp.net submissions -- there must be a better way than the github repo that I have (not) been maintaining :-)

stpeter: Ah, yes. Nick from jabber.calyxinstitute.org has been asking me a couple of times about how to get on the directory.

right

my process is broken

or I am broken

And I think there are a couple of posts to operators@ that have gone unanswered.

yes

exactly

dwd: you're generating valuable content for that blog :-p