small bug who learned to use computers and is evil. 63


QuestForTori
@QuestForTori

I still think it's really important that federated media networks exist because:

  1. It's way harder for a single bad actor (i.e CEO) to fuck everything up in the name of profit chasing
  2. They're useful for weaning people off of the mono-platform hell that the modern internet has been forced into and reminding people how websites used to work and still SHOULD work
  3. Open standards for internet communication can last for ages even with moderate adoption. No one can "take down" the concept of IRC or E-Mail, even if huge mono-platforms have done their best to try and get everyone on a single service for them. You're still usually able to take your data and set up elsewhere using the same backend service, and that's a critical difference. Sure it really sucks when your Masto instance goes under, but it's a hell of a lot easier than losing a proprietary site
  4. Federation makes it much easier to archive content when much of it is already stored across multiple independent servers run by different maintainers

The whole debacle of Reddit's execs burning down the website recently made me think about this more since the same network of micro-communities could be achieved just by a bunch of individual forums networked together with a single-sign-on tying them together. like how forums worked in the 90s and 2000s, just with the option to network them together and not having to make a new account for every community.


QuestForTori
@QuestForTori

I think there are some things that can be done to make federated networks more resilient to individual nodes dropping out, like having each node naturally keep a partial duplicate of some other nodes on the network per the admin's consent, thus allowing for possible failover instances during downtime, etc.

But I think one of the fundamental changes that needs to happen to the internet as a whole is to demystify web hosting and site creation to the modern web user. Services like Wix or Linode kinda help with this, but at the end of the day, they're still proprietary services for a single host. We need a host-agnostic tool with user-friendly UX that can guide a user through the process of buying a domain and/or remote server instance, selecting your website you want to host, or allowing you to select from several open-source server apps that you want to host from an app store-like interface.

We also need better open tools for making websites to go along with that which don't either throw the user into the deep end of HTML coding or lock down so much that you can't make anything other than a bland web-modernist-sludge page.

Once hosting a game server, personal site, or social media node is simple enough for even the slightly-technically-inclined person to do with a platform-agnostic tool, I think people will start seeing webhosting as a lot less daunting.


lifning
@lifning

the potential is there. like, pre-highschool kids who are the slightest bit "good with computers" are out there setting up minecraft servers for their friends.

many computer people, especially webapp developers, love to complicate things and then turn around and say "o, woe! this system is far too difficult and complex for unskilled hands! pray give me the controller, so that i may control it for you." when we need them to say "òwó this system is too difficult and complex! better make it understandable and reasonable to set up so i'm actually empowering people with my work instead of perpetuating the cycle of people-farming"


millenomi
@millenomi

is that it's not a one-and-done thing. It's a time commitment that is:

  • ongoing
  • permanent, as long as the server is up
  • to monitor its usage
  • to respond with manual updates to security issues
  • that scales with the number of components that are needed for the server to run

(e.g. for Mastodon, also with Postgres, with Elasticsearch etc.)

It is a time commitment I literally, a person who has a godforsaken master's in computer science engineering and can absolutely if grumpily look into bootstrapping Linux from scratch if she needs to, absolutely cannot commit to.

And this is on top of keeping up with the social side of whatever activity is occurring on the server itself, if it is serving or receiving materials from or on behalf of others.


lifning
@lifning

right, so the problem statement here is that we want the things-being-hosted to be

  • easier to wrap your head around, such that a greater number of people are eligible to host things, including people with more free time to dedicate, and as a result there are enough people literate in the art of hosting things that the burden need not fall entirely on one person, but on a trusted group
  • less burdensome to the people doing the hosting, such that less free time and energy are required to begin with
  • built for portable identities and data (and with appropriate things encrypted), such that there's less of a commitment of trust and dependency for the user to place on the hoster
  • and crucially, resistant to corporate capture

it feels like the effort being invested into software projects always hard focuses on end-user-friendliness at the expense of everything that it needs to be sustainable


You must log in to comment.

in reply to @QuestForTori's post:

I've been saying (specifically to your point #1) that we need a counterpart to the "Bus Factor" in open source projects, the dollar value for which one can gain enough influence over the service that you rely on to make it a bad place to spend time.

Public corporations have a literal value, based on their stock prices. Private corporations and non-profits are a bit harder to figure out. A diverse group of federated/distributed systems on a common protocol grows hilariously expensive.

ActivityPub and Mastodon both have problems, some deliberate, some short-sighted, and some lacking the attention to fix them, but yeah, people will still be using at least ActivityPub in fifty years, which isn't something I can even imagine for any social media site not run by a government.

in reply to @QuestForTori's post:

We need a host-agnostic tool with user-friendly UX that can guide a user through the process of buying a domain and/or remote server instance, selecting your website you want to host, or allowing you to select from several open-source server apps that you want to host from an app store-like interface.

God I'd be so down to work on something like this if I ever knew where to begin lmao

I don't really know how to solve the broader problem of archival of smaller sites on the internet when they're hosted by individuals who can take them down at any time, other than: It's typically good to have backups across multiple different servers in a network, even if everyone shoulders just a small portion of the burden and the backup software is smart enough to create links for it to find the other content on still-active servers. Also thank god for the Internet Archive, I hope it lives on forever and eventually gets some independent mirrors.

I was musing on this with someone the other day. I'm really big on the idea of localnets on tiny bedroom-level hardware, but even RaspPi availability is in something of a slump or beyond the average non technical user.

I've seen a lot of talk about another ubiquitous computing platform that people have a lot of access to - Old smartphones.

What I think it would take to create a new boom in hosting is something like... you could plug a 5-10 year old phone into a laptop and load an OS that would guide you through configuring a lightweight webserver via the touchscreen, without any prior technical knowledge. A widespread tool like that appears, and you might see a lot more microhosting happening.

ever since twitter started to really go down i've been trying to (mentally) work out a solution for artists to host their portfolio independent from any social media or specific host. there are definitely frameworks to make websites that are really nice to use (astro.build is cool!) (tho, i'm a web dev so i'm probably real biased on this) and there are plenty of hosts that are almost interchangeable between them, some of them even offering hosting for free, with some limitations. as long as you provide the project, it's super easy to deploy with

all the pieces are there for artists to set up their portfolios, all i wanna provide is templates and guides

tho im sure something much more complex would need to be built to host entire communities, i've been feeling this same sentiment for a bit

in reply to @lifning's post:

FUCKING YES.

A huge fucking part of it is how incompetently information tech is taught overall. Impatient "teachers" offended by you not getting it to start with. Jargon being used without explanation, as if you'd already know. Belittling anyone asking on a forum as if they're forced to answer. Long-winded, faux-poetic explanations used over a simple definition. These problems are common to all of academics, but are most certainly not weakened to IT or the internet.

Every fucking time i see a student-blaming statement, i beg of those people withholding knowledge to admit they are a bad educator or stop thinking of themselves in positive terms. Utterly fucking sadistic how much of it is shoved behind paywalls, regionlocks, just general obfuscation.

Education should not only be free, but comprehensible. You want purple prose and expression? Go make some fucking art. If you're educating there's an onus on you, the teacher, to be direct and patient. Otherwise, keep your mouth shut.

in reply to @lifning's post:

the end-user UX is crucially important too of course! stuff has to be approachable on that level as table stakes, or else there's nothing worth putting the effort in to self-host :P but it increasingly seems like people will build a giant overengineered labyrinth of plumbing without enough forethought, make sure the porcelain is nice-looking, and then ship it.

maybe one way to describe what i'm getting at is, when it comes to user interfaces, there're standards and best practices for things like readable text-to-background contrast, time before a page becomes interactable, screen-reader accessibility, etc. out there... but to my knowledge nobody's built criteria or guidelines for the dynamics of "how much time out of my weekends will i lose for the rest of my life trying to keep this thing i set up for my friends working," or "does this open-source webapp document what the social/legal/financial implications of running it on a server are likely to be," if that makes any sense.

at the end of the day it's kinda a UX problem space but rotated a little bit; it's still taking a system and making sure the operator has a mature path to discovering how to understand and operate it, on a bunch of different levels

we have to kill microservices and the notion of scaling, and return to php

i’m not how much i’m joking, but so much complexity in modern web apps can be traced back to it not being enough to just run a pile of code on demand whenever someone tries to load a webpage