@treefit I am convinced that such a bot can do exactly what an app server does in a web app, except it can be run "on a desk" letting the e-mail/chatmail server do the Internetting.
@LorenzoAncora it isn't as well-configured, needs some fiddling around.
Have you ever attempted to write a #deltachat #bot ?
I'm planning to improve bot documentation, what's your feedback about the current documentation state and developer journey?
I should be clear that I admire and respect people who set up Fediverse servers and let strangers join. That is an amazingly generous and thoughtful thing to do. People who take on this and other tasks for running Fediverse servers deserve our gratitude and support.
We are an in-between project. #deltachat is in-between traditional messengers and email apps and #p2p paradigms, #chatmail is in between the massive email server network and #webxdc is in-between e-mail, @xmpp and the web ecosystem (but outlaws central platforms) ... Our approach is rhizomatic, i.e. based on, and changes with, various collaborations and joining contributors. We don't have stable funding and need to regularly apply for public funds ... And are surrounded by VC funded projects
@Ric @matt anybody using the #Jortage service or software to reduce storage duplication across instances? #JustAsking
Thank you @freemo for hosting qoto.org !
@evan
i think a case can be made that torrents have the best track record and are super stable in terms of where the development is. its very straightforward how to seed data , and you can do all of this on your servers so its completely invisible to the user.
downside of torrents is they are immutable . eg cannot change ,, you could chunk your torrent creations possibly or you could use one of
my friend @mauve has an example of how to use mutable web torrents :
https://github.com/RangerMauve/mutable-webtorrent
Thinking about the Mozilla situation, I get the impression that a related problem is that the web world is full of wrong incentives.
E.g. people react to a bad web-site's behavior as "this browser sucks". But since a browser is nothing more than a VM, they could just as well say "this computer sucks".
Maybe good browsers should start displaying "web site benchmarks": for example show in something like a "status bar" the number of HTTP requests performed so far for the web page, the number of hosts contacted, the number of bytes downloaded, the number of JS method calls performed, the number of JS field accesses, etc...
Measurements that try to be browser-agnostic so they reflect the efficiency of the web-page itself.
[ and by "show" I mean "by default", rather than as some kind of developer-only option. So it needs to be sufficiently unobtrusive. ]
Maybe we could then come up with some quantitative estimation of the "useful information content" (UIC) displayed and then show an efficiency result as a ratio of those measurements against the UIC?
Or we could plug average estimated energy cost of each one of those things measured and display the Joules or carbon cost of the web-page?
Another dimension would be to try and collect information about the features used so as to be able to display "displays correctly only with a Firefox more recent than year 20XX".
IPFS no, it's a poor match, particularly for browsers or PWAs.
Mastadon's current implementation is so horrendous it feels like it was specifically designed to enrich cloud providers.
The way PeerTube uses s3 storage, Torrents, Node2Node redundancy, and P2P in the browser solves the issue for video way more elegantly and effectively.
There are much options the Mastodon corner of the fediverse could choose to fix their issues, #jortage is a better starting point for one.
https://jortage.com/ takes the "duplicate every image across every fedi server" model and deduplicates as possible at the media storage level
self-hosting an instance and outsourcing media storage has been for me a nice balance
i just upped my contribution a bit
thought you should all know about #jortage
@mathias I guess the natives too have a reason to be quiet ☹️
@ploum @oneploumshow Personally I'll just consider putting stuff on #IPFS & #BitTorrent instead as that's way more efficient and doesn't require centralized platforms at all...
Just like the ...
@jackwilliambell I had tangential thoughts like this, basically fediverse servers should store posts on content-addressable systems like IPFS, ensure only as much redundancy as they want and reduce storage costs. If it could be done on Bittorrent instead of IPFS, all the better!
pro-libre software, pro-holisticism
pro-communalism, anti-consumerism
fan of #Plan9 and #HaikuOS
anti-witchhunt, see https://stallmansupport.org
I write software (C++) for a living.