Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ibis, a federated Wikipedia alternative

ibis.wiki

Post reply on HN
Print view View original

161–170 of 266 posts

Re: Ibis, a federated Wikipedia alternative

#161

Post by shuntress »

sharkjacobs wrote: post #142

> The fact is that we can’t rely on any single website to hold the whole world’s knowledge, because it can be corrupted sooner or later. The only solution is a distributed architecture, with many smaller websites connecting with each other and sharing information. This article doesn't even try to explain or convincingly make this argument, it just takes it as given.

I would also add that the proposed solution here to the supposed "rot in Wikipedia" amounts essentially to "What if websites? and ... links?".

Re: Ibis, a federated Wikipedia alternative

#162

Post by theamk »

vintermann wrote: post #96
ChrisKnott wrote: post #16

Earlier quoted context omitted.

I just picked a random political article to which I have no significant connection[0]; I don't notice "a lot of bias"...? For me Wikipedia is consistently amongst the least biased sources on the internet. [0] https://en.m.wikipedia.org/wiki/New_Zealand_Labour_Party

If you have no significant connection to it, how would you know if it's biased or not?

In some wikipedia clones, it's pretty obvious:

https://www.conservapedia.com/Video_games

Re: Ibis, a federated Wikipedia alternative

#164

Post by lupusreal »

tux3 wrote: post #133
account42 wrote: post #111

Earlier quoted context omitted.

Wikipedia's notability requirements in particular are quite arbitrary. It doesn't make sens for an online encyclopedia that isn't limited by physical space restrictions to not document everything . If you must, layer curation of what is notable on top of that.

The main concern Wikipedia has with notability is that if you don't have reputable secondary source writing about the topic, then you won't have references for any claims in the article. People just starting adding things that they know, or that "everyone knows". When people disagree about content, Wikipedia always falls back to just reflecting what reliable sources says. And if they disagree, people can always colla…

But "has reputable secondary sources" isn't even close to Wikipedia's real 'notability' determination.

Re: Ibis, a federated Wikipedia alternative

#165

Post by lupusreal »

nequo wrote: post #136
joquarky wrote: post #129

Earlier quoted context omitted.

Perhaps Slashdot had the right idea.

I had to look this up. I think you’re referring to metamoderation where users rate the ratings of comments: https://slashdot.org/faq/mod-metamod.shtml Sounds like a great idea, I wonder why it hasn’t been adopted by other websites. Do you know if it had downsides for Slashdot?

> "Do you know if it had downsides for Slashdot?"

In a word, ossification. It preserves the culture of a forum, to a fault.

Re: Ibis, a federated Wikipedia alternative

#166

Post by outime »

janfoeh wrote: post #80
outime wrote: post #36

Cannot take a project like this seriously in 2024 while looking like total garbage [1] on mobile. [1] https://i.ibb.co/3RWgMGy/IMG-4197.jpg

You believe the interesting, valuable part is the three lines of CSS it will take to improve rendering on baby internet browsers? Not the part where they federated a Wiki-style system?

I believe that if I have a new project and I want people to read about it and get interested, I must support mobile browsers (which nowadays produce more than half of the global website traffic) - more so if not doing it like in my example means not being able to read a single line at all.

It's just basic accessibility what we're talking about here.

Re: Ibis, a federated Wikipedia alternative

#167

Post by mdaniel »

bawolff wrote: post #48

Reading through, this doesn't sound like a "federated wikipedia" It sounds more like they want to implement the github fork & pull-request model of version control where currently Wikipedia uses a more SVN type of version control. There are pros and cons to both models. However federation it is not. The mentioned controversies also seem entirely unrelated to which model you like.

I'm for sure not a federation nor ActivityPub ninja, but as I understand it due to the broadcast nature of ActivityPub and this project's "send updates to the origin server as patches" approach, then it is the opposite of both github and Subversion in that there is not one canonical hostname, and any one hostname doesn't, itself, contain 100% of the content, same as infosec.social for sure doesn't contain 100% of Mastodon content but I can read infosec.social posts on my.mastodon.example by subscribing to feeds that interest me. I currently can't (that I know of) submit patches to their Toots as reply activity, but I don't believe there's anything in the spec that would prevent such a thing, either

I'm keeping an eye on https://docs.gitlab.com/ee/architecture/blueprints/activity_... https://news.ycombinator.com/item?id=39201453> with acute interest because I'm totally open to not even being able to see the ways federation can make ecosystems more valuable due to my lack of hands-on experience with the kinds of problems it is solving

Re: Ibis, a federated Wikipedia alternative

#169

Post by ljm »

sharkjacobs wrote: post #142

> The fact is that we can’t rely on any single website to hold the whole world’s knowledge, because it can be corrupted sooner or later. The only solution is a distributed architecture, with many smaller websites connecting with each other and sharing information. This article doesn't even try to explain or convincingly make this argument, it just takes it as given.

Taxonomy is also quite a complex problem and Wikipedia has put a hell of a lot of work into that, and disambiguation.

The proposed alternative here is to basically distribute the taxonomy? Anybody who's touched microservices for even a minute will know how hilariously difficult it is to co-ordinate not just the services itself but the teams around them, including both the tech and people level ICP.

Re: Ibis, a federated Wikipedia alternative

#170

Post by mdaniel »

kemayo wrote: post #152
ulrikrasmussen wrote: post #27

I'm sorry, but aren't web sites already federated? I don't understand what the technical contribution is here. If we wanted to build a federated network of wikis (do we?), then links to articles are just links and identity is provided by OpenID Connect. What have I misunderstood? > Instead of individual, centralized websites there will be an interconnected network of encyclopedias. This means the same topic can be tr…

I think the bit that makes it "federated" is that they want to use ActivityPub to synchronize articles between servers. Based on https://github.com/Nutomic/ibis#federation I think the practical impact of this would be that it'd be easy to have your personal Ibis instance be a fork of some big mainstream one, where the only difference is that e.g. your version of the Pigeons article explains that birds aren't real, bu…

My experience with the Fediverse thus far also implies that one could run a copy of any such "federated wiki" that has all the same(?) content as other origins but that actually renders on mobile or without JavaScript or via the Gemini protocol or whatever else since it decouples content from presentation. In theory, I would guess that if this also went so far as to incorporate IPFS it could also share the actual storage burden, too, so there wouldn't have to be 4999 copies of the exact same pigeon text (along with 4999 cute pigeon pictures) stored on 4999 different servers plus the one where birds aren't real
Post reply on HN