That's amazing! The forgefriends community (previously fedeproxy) is doing a tremendous job to push for federated forging and break the Github monopoly. Shameless plug for a lengthy read on why that's an important issue: https://staticadventures.netlib.re/blog/decentralized-forge/
It does. Just start calling federated and decentralized services web 3. This stuff is actually useful so it'll out live the block chain pyramid schemes.
as much as I truly despise the current "web 3 is blockchain" trend, I do kinda think "web 3" as "decentralized" works. blockchains are just in vogue (in a big way, to web 3's significant loss) because of a couple unicorns.
maybe this is "web 2.5"? federation brings significant advantages from both centralized and decentralized, while not being fully either.
Which is why I don't think it fits with "web 3 as decentralization", yeah. It's fundamentally different from either, but we lack integers between 2 and 3, and I doubt web 4 would take off.
To my knowledge "web 3" is a very vague term that just means "users host their own content instead of relinquishing it to big tech for hosting". So I think it's fair for federation to count as web 3. Even with a blockchain-heavy net, ad-hoc federation will happen in the form of custodian services.
The top result is Wikipedia, "Web3, also known as Web 3.0,[1][2][3] is an idea for a new iteration of the World Wide Web that incorporates decentralization based on blockchains.[4]"
And even then, with federated services you still usually aren't hosting your content, that's really closer to what web 1.0 was, imo.
I guess the term is still nascent enough for it's meaning to be in flux. Hopefully good people are pushing it to align with the best tendencies
This seriously needs some content. They were accepted to a funding round for an unspecified reason, with a link to a page with the statement "The project summary for this project is not yet available. Please come back soon!"
You should probably check out forgefriends.org. There's detailed monthly reports on what they're doing and why. They started cooperating with gitea because the maintainers were open to the idea of federation (most other forges were not interested) and some french FLOSS NGO (chapril.org [0]) was happy to let them experiment in exchange for maintenance for their forging infra (already based on gitea).
Huge shout out to the forgefriends people! I've never interacted with them personally, but from reading their blog and forums, they seem to take a great care to build a human community before all and to achieve consensus with everyone involved.
[0] It's a pun because it's pushed by the APRIL non-profit, and CHATONS is a campaign started by Framasoft to deframasoftize the Internet (after their degooglize the Internet campaign had "too much" success) to promote hosting cooperatives. The contraction of the two forms "CHAPRIL", a joke a person who is not familiar with the french-speaking FLOSS/selfhosting ecosystem cannot appreciate without context.
Feels a bit like trying to unravel web3 or category theory. I went to forgefriends.org. It's all about forges. Forge-this, forge-that. So like, what's a forge?
Every independent codeberg, gitlab and gitea server (and there are more) that is running is a forge. Right now, you have to have an account on one to open issues, make pull requests, contribute in any way to code hosted on it and only it, and you need an account on each and every server you want to contribute to, even provide a bug report. This is no good, because people who like to host their own code can only frictionlessly contribute to their own code, or code hosted on their forge, or the forge they put their code on. How many people are going to create an account on gitea.bobsserver.biz just to put a bug report in for an annoyance in some app they use?
This is why github has a near monopoly in this space, everyone can contribute with a github account. To decentralize this, I need to be able to put a PR in your project hosted on your forge using the account I have on my forge. This is what ForgeFed does, it is an extension of ActivityPub to enable federation of forges.
Documentation continues to be the most undervalued and underappreciated part of open source while being absolutely essential for any project that cares about adoption.
Isn't the problem of having universal accounts across servers independent of git and hosting code? I don't see how the two are related.
It's a good goal, b/c at the moment you only have "Login with Google" or "Login with Github (ie. M$)". But I hope the ultimate solution isn't compartmentalized to source code management
Maybe I'm not seeing some aspect which ties the two together?
At present, you can move a repo between forges by cloning it locally and pushing it back up. However, you lose the issues/milestones/PR comments etc.
Federation is useful since it defines a portable standard for those things.
If you have an account on federated forge site A, you can use it to file bugs (via the forge A UI) against a project on federated forge site B.
Both websites UI will show the bug and any comments on it - data will be replicated.
In practice, there'll likely be complexities around EG moderation (particularly for larger projects making controversial decisions), replication occasionally losing data due to intermittent faults, etc. IMO those technical reasons are less likely to sink the project than inertia is; people are already using github.
I don't think inertia is really a problem. People are already using these self hosted options, and particularly in FOSS, people value the option to be independent. Maybe it won't sink github, but it will probably end the practice of people mirroring there and running their own canonical repo elsewhere.
I went to the site and the second line on the page says “Because most Free Software projects are hosted on proprietary forges (for example GitHub)”. So code storage and collaboration.
A forge is a location people meet to build tools. Except we don't melt metal parts in a big fire on Gitea/Gitlab/Sourcehut so the analogy runs dry pretty quickly.
What's basically going on is that a project to extend ActivityPub (the protocol that Mastodon, Pleroma, Peertube, Lemmy, etc use to federate with one another) has a proposed extension called ForgeFed to enable more interaction types so as to allow code forges to federate with one another. There have been a few projects slated to implement it once the spec is finalized, gitea being one of those projects.
Any activitypub experts out there?
I have some interesting use cases with adding activitypub support to allow an app I built[1] to communicate across instances. Would adding activitypub support allow it to integrate easier with other activitypub apps too? Thinking Nextcloud etc. Can it replace OAuth possibly?
This thread seems to bring out the worst in me, but what point is there to a "federated" closed source app?
What a closed source application implies is: I want to get access to all these people in the fediverse, yet I am not willing to do anything to give back for that. To me that sounds like a bad deal, for any self-titled "experts" out there.
That's a fair point and I can't argue with it. Can you email me (contact in profile) sometime? I have some questions for you about closed source and what would make it a better deal.
Sorry but I'm plenty busy with my own projects (you can see them in my profile). The ActivityPub packages have a mailing list, feel free to ask questions there if you see anything relevant for you.
> The Fediverse (a portmanteau of "federation" and "universe") is an ensemble of federated (i.e. interconnected) servers that are used for web publishing (i.e. social networking, microblogging, blogging, or websites) and file hosting, but which, while independently hosted, can communicate with each other.
Usually they talk about ActivityPub, as a growing number of services are based on this (or similar).
I was honestly confused for a second why would you think the name had anything to do with a fetish (minus some mastodon content) until I realized you saw it as fe-diverse and not fedi-verse. I never saw it that way because I pronounce the "i" in it as "ee", not as "ay". I wonder what's more common now. Is this another gif/jif situation?
Apologies for the negativity, I'm feeling excessively bitchy at the moment, but how about you improve your reading comprehension instead of having bad takes at other people's naming efforts?
It refers to the domains federating via the ActivityPub protocol. Software or deployments of software able to talk to each other via ActivityPub are referred to as the Fediverse.
No it doesn't, the tech crowd settled on fediverse, make your peace and move on. It's might be dumb sounding term, but then so was "electronic mail" or "wiki".
Despite your framing, I'm actually curious about your criticisms. To me, a federated world is strictly better than the increasingly centralized world we find ourselves in. Are you suggesting anything less than fully decentralized is not worth the effort?
I'm being fairly literal when I say that federated social networking is feudalism.
That's not a better/worse value judgment -- just pointing out that some folks make the mistake of believing otherwise.
As far as value judgments go though, I've long been saying that the current model for the internet is crushing us; that collectively as a society we should get together and abolish the use and creation of social networking sites (the same way qualuudes don't exist today); and that "skin in the game" should be required for online services -- no more "free" stuff where the product is actually the user. I will gladly pay for an email account that isn't mining my data, or a search engine that's actually useful, etc.
Thanks for the clarification. The part I disagree with is the binary classification, and also individuals paying is not a guarantee to get free of the power structure, for instance take Youtube premium. It's not obvious that the platform itself must concentrate power, but I totally agree that the power differential is core and needs fixing.
its a maximalist position that is hard to parse. you can run your own instance, keep it very local to a group of friends and family etc. what is feudal about that?
Does this mean you will be able to fork, pr, browse history/diffs, and network graph across sites?
MS/github may successfully keep this from gaining popularity through the usual means, but to me this is huge and should be huge.
If the interoperability mechanism is open, maybe gitlab can join it. That might have some decent effect.
3d printing needs this same thing. Or rather than some new federated thingiverse, existing git front-ends in general really only need just a little bit extra and they could support open hardware as well as they currently support software and even better than current hardware design sites. It's really just down to presentation. You can almost do it now in existing github. It has a 3d model renderer that can be embedded in your readme and even a stl diff viewer.
(Though, I consider stl's as compiled output and placed in release assets, not source, so an stl diff viewer essentially doesn't apply. I guess that exposes the problem right there. A code diff viewer is practical because the code itself is directly viewable and meaningful to a human. The source files of most cad and pcb apps is not, so you'd need 100 different renderers of 100 different commercial products to display meaningful diffs)
> Github is a social site foremost, a repo host second.
Hopefully EU regulators will take the same view if the Digital Markets Act becomes law.
"Interoperability measures were another major addition to the Parliament’s text, requiring social media and messaging apps to communicate with each other."
Maybe cooperability is already a solved problem? As I understand, Linux kernel changes are managed entirely by text-only email, and email is of course entirely forge-agnostic.
Yes, it is solved in this particular context: Linux kernel, Qemu and all Free Software projects that have a mail based workflow. But there are other contexts where Free Software developers chose a different workflow such as using a forge. Whatever the reasons, this is where we are now. There are two paths forward: a) convince people using forges to switch to a mail based workflow, b) federate forges. I believe SourceHut and many other do a fine job trying to convince people to switch to a mail based workflow. The https://forgefriends.org and https://gitea.io projects chose to try to federate forges. It's not a competition, there is value in both approaches and they will apparently move forward in parallel for the foreseeable future.
"Becoming a social site" is easily taken the wrong way, it doesn't mean "becoming like facebook."
A big part of FOSS is community contribution. This is social in nature, fundamentally. So forges are already social in the sense that many people can contribute with bug reports, PRs, comments on issues, forking, merges, etc. These forges wouldn't be becoming more social, they're already social, all that's changing is they won't be islands anymore.
+1 for gitea. I've messed with codeberg, sourcehut, gitlab, even fossil (non git version control system, fantastic software but unfortunately everyone uses git) and I've got to say gitea is where it's at. It's the right combination of light and feature dense, could be lighter IMO but it's great.
They're not providers, although gitlab and sourcehut do offer code hosting services, and I'm sure some people are offering paid access to services they host with some of this software.
Codeberg could be said to be a provider. They are hosting FOSS projects exclusively and contribute to the Gitea project, as that is the forge software they use. They have services like Codeberg Pages and (soon) Codeberg CI.
I've been using Gitea for a while now, and have had no issues. I self-host it for personal projects, and also use it at work. I would absolutely recommend it.
The way I see it, the choice is pretty risk-free. I currently mirror repos on repo.or.cz (the original git host), codeberg.org, tildegit.org and sr.ht. (Feel free to suggest more!) I use pushurl extensively to transparently keep them all in sync: https://stackoverflow.com/questions/14290113/git-pushing-cod...
Over time it'll become clear if some are unreliable or have other issues. When that happens you'll be glad for the diversification.
This whole evolution reminds me of the early web, which was often a thin veneer for multiple ftp mirror sites. This makes me happy :)
I don't have comparisons between the two but I'm a sourcehut early adopter paying at the $100/yr level and I'm more than satisfied for my personal projects.
Drew's head is in the right place and nowadays that matters to me for service providers more than anything.
Forking and merge requests across instances. Also things like retweeting merge requests and comments from Mastodon, and having Mastodon replies appear as comments on the Gitea issue/MR.
There should probably be a way to merge project lists on various forges into a potentially huge index that can be browsed and searched, independently of each particular forge. Maybe it already exists?
Federating software projects so they can be interacted with from any forge frontend is the main focus. But federating the search engines internal to each forge is another important aspect. When federated search becomes a reality, looking for a project B on forge A will give an answer even when it is hosted on forge B.
Nowadays, with over 90% of the search run by Google, it does not matter much. Discoverability is whatever Google decides it is. But this is not desirable and more importantly it should not considered to be a solution to the discoverability problem. On the contrary dominance of Google is the main problem of discoverability and the only way to solve it is to provide sound alternatives.
imho the fediverse will have to come to be much more than social media type exchange of messages via activitypub. federated search is a major additional pillar that is currently largely absent (and is technically very different)
Well, there's sepia search for Peertube, possibly something like that could be built into a forge server software one day. I don't believe it is in the ForgeFed spec though and IMO it doesn't belong there.
please publish DOAP records, so as anybody can see where things reside. It's easy and launchpad has done it decades ago already. Now federation would complete.
Worth pointing out that activitypub, the fediverse etc are the actual 'web3' decentralized systems people seem to be so hyped about, with the added advantage that they run on normal, fast computers and not on the equivalent of a car driving with the breaks pulled so someone can mine coins.
Well, not quite. They're definitely an improvement to centralized silos, and there's a prominent place for AP in the future of the web. But a big problem with the spec IMO is that it relies on domain names to identify servers.
To really rake things to the next step, we need public key addressing. Ubiquitous adoption of IPv6 would suffice, but public key addressing would be better. Either way, using a domain as a permanent identifier is a fragile and brittle.
Another big problem with AP is that the spec is extensible in ill defined ways while being designed with current social UX in mind, thus counterintuitively making it needlessly restrictive. So there are interoperation problems with different server software that implement the spec differently (look at federating servers like peertube, Lemmy and mastodon, while possible, it is clanky), meanwhile it has to be modified to support ForgeFed (which is what gitea will be using to federate forges).
how is ipv6 addressing not worse than DNSes? Like you move to a different house and suddenly your ipv6 address gets lost?
Perhaps your argument is around DNS spoofing or something but domain names feel pretty non-controversial for most usecases, especially for the fediverse, which is about federation. There's an expectation of multiple multi-user services being the norm!
Domain names are very, very useful. But with regard to activitypub, if a server wants to change its name the federation and social graph is broken entirely. Even if your server goes down and needs to be rebuilt, federation gets broken due to this design decision, because other servers don't know you're offline and new status messages get lost.
I guess “domain name change” seems like a much more manageable thing than “IP address change” (I don’t believe IP addresses are portable between hosts?)
I think I understand better what you’re saying though
Well so if servers were able to identify themselves to each other using a public key, they could change domains whenever they wanted and just update the others with a signed message. You'd still use DNS to navigate, for user identifiers, but a change wouldn't break federation. Unfortunately this is not in the spec and it is a bit involved to do.
Well... no, nobody is really hyped about them, but I do think as a technology it is more worthy of the web3 moniker than anything in cryptocurrency land. I like the fediverse, and the user experience is decent, at the same time, there's not really a great reason to branch out and use it if you're living in twitter land, the product is basically at parity with twitter, it needs something to differentiate it (for the masses).
fediverse > mastodon. granted the other platforms not yet well populated but if you look e.g. at bookwyrm, its got nothing to do with twitter https://bookwyrm.social/
Good for them, but I (as a EU citizen) don't care about projects like this.
It is a soulless copy-cat of a much more famous, profitable, free and useful project.
Just look at the logo, what is that line in the middle and why is its stroke with bigger then the other stock icon embedded in the icon.
I know I'm knitpicking but it says something about the character of the project and NLnet. This is woke IT, as long as it is decentralised it is somehow better than other software no matter how crapy the UX.
And please don't tell me it is open source software and I can contribute something, like I have time for that. Just take my money and shut up instead of, you know, using my tax money for stuff I don't care about. Thanks for listing to my TED talk.
Is it really soulless? It seems like it solves a real problem: many Gitea instances are very small. The devs are passionate about their work and they are all working toward improving the software. So what if it's a copycat? GitHub could be argued to be a copycat of SourceForge. Gitea/forgefriends offers new features and improvements over GitHub. Gitea is as useful as GitHub for hosting a git repo with a web interface, but the federation aspect makes Gitea also a social site (much like GitHub).
As for the logo, the line in the middle is the "string" for the tea bag. It is a different size because the actual logo is not part of the string, so the distinction must be made.
> This is woke IT, as long as it is decentralised it is somehow better than other software no matter how crapy the UX.
As someone who has used Gitea in the past, it is wonderful for it's goal of running a Git server with a webUI. Don't knock it until you've tried it.
> And please don't tell me it is open source software and I can contribute something, like I have time for that. Just take my money and shut up instead of, you know, using my tax money for stuff I don't care about. Thanks for listing to my TED talk.
If you're not willing to work toward fixing something, then stop complaining about it.
I paid $0 for about a year and a half on Vultr since they had a promotional $50 credit. I got their cheapest option and just added some swap space. Plus, I like hosting my own compared to centralizing everything in a Microsoft platform
Your apparent obsession over wealth redistribution vs individual property seem to give you a peculiar view of the social universe. I don't believe we can have a meaningful conversation.
EDIT: Apparently there's a "forge federation webinar" coming up soon: https://forum.forgefriends.org/t/forge-federation-webinar-ja...