| 12 Dec 2018 |
| 20:33:11 |  neilalexander | I suppose a not terrible solution would be to get Yggdrasil to "cache" metadata for hosts and if it finds some kind of SRV data then it could beacon some DNS-SD packets to the host on the loopback adapter or something |
| 20:33:13 |  yakamo | im thinking if a user finds a serivice or something they can it them self |
| 20:33:24 | yakamo | if a second is found it gets ammended with a num or something |
| 20:33:30 |  Arceliar | neilalexander: i'll take a closer look after work, but 99% sure that's getting merged tonight, so we should at least have develop builds that don't depend on systemd right away |
| 20:33:30 |  TimMc | DNS is what you get when you take Zooko's Triangle and decide to toss the "secure" axis. |
| 20:33:55 | neilalexander | Arceliar: Yeah, cathugger was quite vocal about it on overchan |
| 20:34:19 | yakamo | lol |
| 20:34:21 | yakamo | thats fair enough |
| 20:34:43 | yakamo | at the risk of getting kicked from the channel, i dont mind systemd so much at the moment |
| 20:35:20 | yakamo | the startup units are a vast improvement |
| 20:35:26 | Arceliar | TimMc: i'm of the opinion that the "human readable" part is the best one to drop. humans don't use the network, computers use the network, and they're going to outnumber us in the long run. that and, lack of DNS encourages distributed/p2p applications over centralized ones. and bookmarks exist. |
| 20:36:26 | Arceliar | it's not a popular opinion |
| 20:36:40 | TimMc | It's hard to put on a business card. |
| 20:36:48 | TimMc | You're basically reduced to using Google as the DNS. |
| 20:36:59 | TimMc | AOL Keywords, 2018 edition |
| 20:37:15 | Arceliar | a QR code isn't terrible in that scenario |
| 20:37:19 | yakamo | could put the address in a qrcode |
| 20:37:23 | yakamo | lol |
| 20:37:25 | TimMc | Yeah, true! |
| 20:37:48 | Arceliar | QR code is even easier than typing if you're on a phone. and once you get there once, bookmark. |
| 20:38:08 | yakamo | i like that idea |
| 20:38:24 | Arceliar | also, lots of people don't seem to type in URLs, they google search them and then click the first not-obviously-an-advertisement link. not thrilled about that, but it is what it is. |
| 20:39:02 | TimMc | I like your optimism that they manage to avoid the advertisements. |
| 20:39:39 | Arceliar | i did it was only obvious advertisements they skip... |
| 20:39:44 | TimMc | true |
| 20:40:07 | TimMc | I've been pondering a dual system where there are non-human-readable names (like IPv6 addresses), and then non-unique names that get attached to them via a reputation system. |
| 20:40:49 | TimMc | If site 123 links to site 456 and *calls* it Wikipedia in doing so, and you trust site 123, your client can record that association automatically. |
| 20:40:54 | | * yakamo wonders if TimMc is chinese |
| 20:41:18 | TimMc | hah, not that kind of reputation system :-X |
| 20:41:25 | yakamo | :) |
