Notices by mike (mike@loadaverage.org)
-
mike (mike@loadaverage.org)'s status on Monday, 26-Sep-2016 00:46:57 UTC mike The way that these W3C social initiatives have played out is that they start by mandating a protocol to solve everybody's problems. Everybody rubber stamps the protocol and then we start building applications based on that protocol. What they're missing is the first step - defining the problems that need to be solved. So far we're on about the 5th or 6th attempt to define a protocol to send messages back and forth. We've been sending messages back and forth since the 70s - it isn't that hard. Defining how you interact with privacy and permissions and *differences in permission models* is bloody hard. Every project has a different permission model. This is where the W3C needs to start if they ever want to be taken seriously. It's entirely possible (as we've discovered) that the standard protocols don't cut it when faced with these requirements. You need a different protocol which is desig… -
mike (mike@loadaverage.org)'s status on Sunday, 25-Sep-2016 20:43:23 UTC mike Unix has a 'public' (e.g. 'others') bit. Are you suggesting that Unix should scrap this or make everything public; because of the conflict with trying to provide working permissions?
Yes, you certainly need to work harder on your security when you mix public and private in the same system. That doesn't mean you have to limit the functionality. You just need to fix the bugs and over time bake the permissions into the 'kernel' of your app so that it is harder for kiddies in userland to screw them up. -
mike (mike@loadaverage.org)'s status on Sunday, 17-Jul-2016 12:56:16 UTC mike Friendica, Diaspora, and Hubzilla all use a similar mechanism where comments are sent upstream (analagous to the intent of the salmon protocol) to the conversation root and then redistributed downstream to all the original participants. The end result is that all of the original participants (or recipients of the original top level post) receive a complete copy of the entire conversation. GNU-social uses a slightly different distribution model where there is no distinction made between upstream and downstream distribution. As a result the conversation distribution can vary depending on the relationships of the nodes involved. -
mike (mike@loadaverage.org)'s status on Tuesday, 05-Jul-2016 06:43:19 UTC mike Hubzilla respects privay and doesn't try to manipulate your feed to sell you stuff. It is also open source so you can inspect it and see what it is doing. So no - it doesn't imitate Facebook at all. -
mike (mike@loadaverage.org)'s status on Thursday, 03-Mar-2016 23:29:45 UTC mike @pipistrellum There are several sites which don't require invitation. https://hubzilla.site and https://lastauth.com should allow registration. There's a list of open sites at https://macgirvin.com/pubsites but be aware that macgirvin.com itself is a private site. -
mike (mike@loadaverage.org)'s status on Monday, 29-Feb-2016 01:39:05 UTC mike After a few false starts - the hubzilla "pubsubhubbub" plugin seems to finally be pushing posts; so you should be able to follow *public* hubzilla channels/threads on GNU-Social.
I'm expecting you will encounter an author attribution bug on wall-to-wall posts and comments which was reported to statusnet years ago and I believe still exists. Basically if you get an Atom ActivityStream and the item:author is different from the feed:author, the item:author is ignored or discarded and the activity is attributed to the feed:author or identity that you are following - and this is wrong. So it looks like everybody you follow is talking to and even arguing with themselves. -
mike (mike@loadaverage.org)'s status on Wednesday, 28-Oct-2015 01:50:01 UTC mike #hubzilla does not federate nearly as well as #friendica. Federation and nomadic identity aren't rea…