Conversation
Notices
-
Thomas Willingham (beardyunixer@soc.beardyunixer.com)'s status on Thursday, 18-May-2017 22:20:46 UTC Thomas Willingham The problem with activitypub is really quite simple. I don't see why folk are having a hard time seeing why nobody is very interested after the brief initial burst died down.
The two major concerns I have with the protocol I'm using now are that a) my crypto is cryptographically broken, and b) I don't have envelope privacy.
The difference between the protocol I'm using now and ActivityPub is I'm already using the protocol I'm using now.
You're asking me to implement a new protocol that doesn't fix the problem I already have, and guarantee an overall loss - as we'd lose years of testing and bug fixes, while introducing new ones to have the same broken elements.
The rest of the argument doesn't matter. It doesn't matter that there are regressions as well, I don't need to argue that far - you're simply not fixing the problems we already have, so there is no reason to change.- Claes Wallin (韋嘉誠) repeated this.
-
Thomas Willingham (beardyunixer@soc.beardyunixer.com)'s status on Friday, 19-May-2017 18:04:45 UTC Thomas Willingham And what do you mean "nobody is interested"?
I'm seeing much more opposition than enthusiasm in my stream.You're just bashing it
Not really. I basically said "I have these problems and so do you, you've got nothing I haven't already got, so why should I bother?". That's hardly bashing. When I'm bashing you, you know about it.the privacy envelope I get, assuming you mean private payloads are not encrypted
Not the payload, the metadata.suppose Hubzilla supported sending and receiving stuff using ActivityPub, and we started to send our assets to other parties, e.g. a Zot app or webpage or a wiki - what are they to do with them? Is there a way to tell what to accept and what to do?
It's extensible, but in the same way activity streams are extensible. These are already activity streams, and they'd perform as they already do (ie, not well).Claes Wallin (韋嘉誠) repeated this. -
Thomas Willingham (beardyunixer@soc.beardyunixer.com)'s status on Saturday, 20-May-2017 21:35:42 UTC Thomas Willingham Doesn't mean we should try to be compatible with each other.
99% of fights are because everyone is upset their homozygous monoculture isn't the one accepted by everyone else. Everyone wants to be compatible, but they want everyone else to be compatible with them, not the other way round.
A W3C standard that actually solved the problems of federation - which are nothing to do with the protocol in the first place, the protocol is the easy bit - would need to be extensible, not compatible.
But it'd have to be really extensible, not this wishy washy "you can create new activities, and then nobody will understand them!" way we've had for a decade. We need a way to transmit server capabilities, and we need a way to describe what to do with an item if the recipient is not compatible. At a minimum.Claes Wallin (韋嘉誠) repeated this. -
Thomas Willingham (beardyunixer@soc.beardyunixer.com)'s status on Saturday, 20-May-2017 22:11:26 UTC Thomas Willingham @bob I don't know how that works. I only really used XMPP as a, well, user. But maybe they do already do it. It's not far fetched.
When I send [flibble]Some new flobble[/flibble] all we need to solve 90% of arugment is an extra bit that says "If I send a flibble, post a url to my flobble". That's it.Claes Wallin (韋嘉誠) repeated this.