XSF Discussion - 2015-10-16

Ge0rG: inline signing, s/in/on/ ?

Haven't seen inline before in this context, but it sounds like you mean online signing

Zash: Well the BIND feature really appears to be called "inline-signing"

I think PowerDNS calls it online signing

Ge0rG, That blog post looks good. I might have to try it all out; there's a free DNSSEC anycast slave service run by a guy I know that I've been meaning to use.

"I know a guy"

Exactly.

I have been using inline signing in BIND for quite a while now. Very useful.

I do offline signing, in which the signatures are embedded in the zone file. Almost like they are inlined. Imagine my confusion.

Kevish, Been thinking about PEP, MUC2/MIX, and MAM, BTW - I think we need a way to distinguish between ephemeral and whatever-the-opposite-of-that-is events. I think MAM currently distinguishes these purely on the basis of containing a <body/>, but I think the reality is more complex than that.

Not sure how you would use a slave server with online signing

dwd: Stanza profiles?

Possibly, but I think it might need per-event processing. If MUC2/MIX sends real message stanzas (as opposed to pubsub events) it's simpler, but otherwise non of those will be captured by MAM.

dwd: Could be.

I kinda wanted to use type == chat or has body

Zash, I would prefer as simple a rule as possible, as well.

Currently enjoying my archive being 90% chat states

hehe

But isn't that useful?

I'd have thought it'd be important to know that before they sent am essage, your contact started typing it.

Useful detection that your contact isn't a bot, for example.

I'd like to say in the mam query 'only messages with a body, please'

Zash: Nothing stops your server offering such a thing.

Given the data forms.

Zash, Well, that's another issue - filtered MAM queries, since you want to have clients pull, say, geoloc events as well as chat.

Yeah, we've got that already.

Just shove whatever you want in a data form.

Kevish: my laziness and desire for a simple implementation does

And then, if necessary, standardise the form fields.

Kevish, I'm not claiming it's difficult to do.

Kevish, Just claiming that we'll need to address these things.

Right, I was saying that we have addressed 'these things' as much as we need to generally.

And it's just a case of speccing (or not) each individual thing that needs handling.

what about things like pubsub notifications with body elements for clients dat don't support pubsub?

I've seen useful applications of that, using type=chat, too, by the way.

ralphm: I think that in this day and age that's a bad thing, because it's going to mess with so much.

(I suppose I'm not opposed to MAMing pubsub, but hey, I wanted to bring it up)

(Outside closed systems that can do whatever they want)

ralphm: PubSub has stuff for configuring that right ?

Zash: the protocol might have. Implementations, I don't know.

There is, of course, the <no-store/> hint.

Zash: re DNSSEC online/inline signing, you can also run a hidden primary that does the signing...

Ge0rG: I could, I have two slaves.

Zash: sex slaves, I hope.

Inappropriate.

I'm sorry. I'll guess it's my head when the robot revolution comes

Kevish: sorry. I tried hard to come up with a good SJW argument and failed.

The menu on new.xmpp.org doesn't appear to work on my phone

Dear HN sock puppets, ehrm... I mean, dear proponents of XMPP on HN, please contribute to https://news.ycombinator.com/item?id=10400184 / http://op-co.de/blog/posts/yax_im_dnssec/

'dat HN title...

SamWhited: it's the post title, and HN policy is not to edit titles.

Is it? I feel like most HN posts have a descriptive title... interesting.

SamWhited: probably because the original titles are descriptive as well.

Looks like people are playing techno buzz bingo on HN

intosi : yes they are.

Given that the article has 'I' in there, who wrote it? Why isn't the author's name visible near the post?

Ge0rG: (I assume it was you)

ralphm: I did. It's on my personal blog, which is based on ikiwiki, which is rather arcane.

Ge0rG: you can't get an author's name in there anywhere, but have funky titles with mixed markup?

ralphm: there is no mixed markup in the title. the long version of the title is merely bold text.

Oh, I thought the subtitle was a separate thing

Just submitted a new protoxep for something at work; if anyone's an XML schema expert, would you mind reviewing? I just copy/pasted from another spec and changed the namespaces/element names; no clue if it's remotely correct: https://github.com/xsf/xeps/pull/110

more direct link: https://github.com/hipchat/xeps/blob/attach_to/inbox/attachto.xml#L156

Schemas for Experimental are optional.

Oh, that makes sense; shouldn't have bothered. Oh well.

Saves you time later.