Here's my thoughts on how we can replace the web browser web with a data web and how this Reddit Migration thing is a great start. (thread)
One pain point though is exploring the domain specific way of formatting data while still being able to do all the social stuff. E.g. right now if I want to see a lemmy thread, I need to either see it boosted or navigate a lemmy instance and copy paste a link.
This means having lots of apps and identities, or also jumping between a bunch of websites.
This is where we are with the web and the current state of apps. A bunch of silos and we let them take full control.
However! This is also where the SemanticWeb and ActivityPub fit in.
One grossly underuse feature of activitypub is the client-server model and having the option to have access to data using a proper JSON-LD interface.
In other words. What if "fediverse" clients were just "Semantic Data" clients. Instead of relying on a specific app, we could have views over the raw data within a particular community through common (and swappable) interfaces.
Here's how we can get there and why it makes sense
One thing that we have now semantic-web-wise is a lot of data that random folks care a lot more about (it's their community or special interest!). And this is also coming from folks that were likely using alternative APIs already so they could get the best view of the data for their use case.
Usually this means new apps for specific purposes. But what if we could simplify making those apps and being able to recombine elements between them. In some sort of "user agent" for example.
Instead of a bunch of adware and megabytes of JavaScript / developers implementing totally random ways of doing things sites can serve "data" directly. Your browser should then take that data and give you some nice way of viewing it or letting you look up ways of viewing it (or buying or whatever for people that want that).
We could already start seeing things like that now with the ActivityPub data being published out there. What if the data itself also had suggestions for "default views".
Here's some things that I think would be important to have:
- User configurable views of the data (with color schemes!)
- Registries for data types and community curated lists of views
- Find ways to pull most common data on the web (e.g. reader view parsed html as legacy web page viewer, fediverse data)
For User Configurability, I think that's one thing that makes me the saddest about the state of the web (and apps!). While some hacker types can inject stylesheets and edit the layout at our whim, the average person is basically stuck with all the banner ads and useless shit and inconsistent accessibility.
For me personally, a lot of sites just don't work for my needs and are a constant pain to try to read through. What if instead they used standard ways of presenting their data.
And in order to make it easier for people to interact with data, there should be ways to have at least a default or some sets to choose from.
From there people could write their own views and share them with their pals or within moderated communities of view tied to data types.
Like, right now I have no clue what apps forlks are using, but it'd be cool if someone I'm mutuals with got added to my trust graph and I could discover the stuff they think is cool.
This gives space for creativity
So, I think the SemanticWeb vision has probably been around there for years. But what's different now is that we have a bunch of data for folks to be into! If we get even a rudimentary thing that could let you start exploring the different fedi data types and let you bring your account with you to be able to comment whenever you want, we'd already be serving a bunch of folks.
I'm personally that folks too. :P And we could also be making a new foundation for a user-centric web.
I think one thing that would be crucial on top of starting to get new views, would be to normalize the use of Triple Data Fragments on any ActivityPub and SmanticWeb data. Without even the most basic filtering on datasets we stand to have a lot more waste and UX problems . If I need to parse a person's entire history every time I want to view their data, it's just going to be too slow.
If clients could fetch just the "latest" few posts from a standard AP source, we wouldn't need servers so much
@mauve
this is literally what I'm working on rn and also it is good to see another person being like "say we should actually use the JSON-LD part of activitypub"
@mauve Have you seen what Arc Are doing with custom style sheets yet ?
@RevPancakes Yeah! It's really cool what they're doing with moving snippets of pages around. I kinda wish they didn't require a proprietary cloud service to work though.
@mauve I’m with you on that, it would be great if we had similar things in floss browsers that are not at the whims of the business model and continued existence of the ‘browser company’
@RevPancakes I'd love to explore some of this stuff in Agregore but sadly I have been too busy with client work to spend much time rethinking it. Just been slowly ideating and adding lil fixes when I have time
@mauve 👍
tbh the @w3c #socialwg charter kinda reads like the group should have done more on 'views'
opensocial had 'preferred experiences' https://opensocial.github.io/spec/2.5.1/Core-Gadget.xml#preferredExperience
@mauve @bengo @w3c Hope so! Does feel like the #RedditMigration has introduced a disruption in the ether. 🤞
One thing that we're seeing with stuff like kbin and lemmy is that there's room for other formats of the fediverse that can be appealing to lots of communities. And with AP at the core we can at least get some basic interactions server to server regardless of the specific format.