Add Explainer

[michielbdejong]

As discussed this morning, for the TAG review, and for anyone else looking to understand our charter, we should add an explainer that starts from a user perspective and explains what we’re trying to build and why.

CC @pchampin @laurensdeb @acoburn

[michielbdejong]

“What is the user need that is being solved?”

And mentioning the storage server doesn’t talk directly to other storage servers, although e.g. ActivityPub could be added as an outgoing notifications channel type, or a bridge server (proxy) could translate e.g. WebSub notifications to ActivityPub ones for consumption by SocialWeb nodes.

[michielbdejong]

The protocol for blog authoring tools we talked about is this one: https://en.wikipedia.org/wiki/Micropub_(protocol)

[elf-pavlik]

I don't understand how Micropub should be related to the Solid WG charter. Does the charter need to clarify that Solid Storage is based on REST rather than RPC?

[melvincarvalho]

Isnt micropub a competitor to Solid for people that dont like RDF?

(scroll down to see selected comments)

https://indieweb.org/RDF#provokes_emotional_ranting

[melvincarvalho]

“What is the user need that is being solved?”

Solid offers a Turtle API with auth & some other functionalities. While its technical advantages might not be immediately clear to the general user, it can hold significant value for those researching the semantic web and for those developing linked data applications.

Perhaps a better user-oriented question might be:

What unique problems does Solid solve, which cant be done already elsewhere?

[woutermont]

I'd refer to @sandhawke (source, my boldface):

... much of the motivation for solid was to allow people to move easily between competing apps. That was something we didn't know how to do even if we wanted to, when the apps were social. There's a kind of lock in, Social Lock-In that gives that software a huge barrier to entry against competition. So, in that sense, most of the advantage of Solid for humanity comes from its use in Social computing.

Solid's promise is indeed advanced interoperability, separating data from applications, making data ultimately portable. It promotes universal authentication mechanisms and combines it with universal APIs. It is not "one API to rule them all", but an API standard that allows innovative development of and competition between data interfaces.

[sandhawke]

Thanks for the tag. Fwiw I'm calling that property "App Freedom" and recently gave a talk about it, linked from appfreedom.org. Slides are there, too.

[melvincarvalho]

@woutermont @sandhawke great points, I watched the video and enjoyed the slides. The site was slightly low on details, but the hint was something like solid but with a better DevX (JSON + schemas). That's seems better than solid's universal Turtle API, for modern web developers, I think.

I'm going to be slightly devil's advocate here, so that we might be able to get to some good text. Let me look at the benefits of solid mentioned through a lens of USP:

• data portability, talked about since 2007 (Data Portability Group) we tried it in crosscloud, migration is tried in mastodon. It doesnt really work with the federated model because you might change internal links but cant change external links. We're always 1-2 years away from a half solution. Case in point, my data was migrated from one server to another about 4 years ago, and to this day the new server doesnt display my webid correctly which I put a lot of time into, and I didnt even ask for the move. It would be nice if it happens, but Solid doesnt promise this today.

• advanced interoperability. Does this mean a universal RDF format that is based on turtle and RDF? Or is it a JSON format with loose schemas that sandro alludes to? Does this even fully work, e.g. with arrays? As sandro once said to me, RDF doesnt guarantee you interop. It just gets you to the edge of the abyss. (I hope havent misquoted you there). I do personally think RDF has a lot to offer. But unsure it's captured in the phraseology.

• universal authentication mechanisms, im not sure on this one either. Lots of things you can log in with OIDC. Solid was really nice for me when we could log in with TLS, and I'm a bit sad that has been commented out. Is this really universal authn. Maybe it's a kind of industry standard. I actually struggle to log in to any solid instance at all these days. Maybe that's just me, because my webid is on my homepage, I'm prepared to write an OIDC provider and run it, but there seems not module for that.

• universal APIS. Could be onto something here. But if it's mandatory Turtle, again, that's a tough sell. What is the universal API. There could be a way to phase this in a compelling way.

• so then we have competing apps. I think this is a good pitch. Though it is done elsewhere. However, when was the last time you used competing apps with solid and had an "ah ha" moment. Something that provided a lot of utility. Normally I write my own apps, so I can see that. But do folks actually use it today, or is it still a promise. I get maybe there could be different file explorers, or profile editors, but we have that with dropbox etc. Does solid have a killer app that people use alot, with an alternative?

Maybe it's a combination of all these things. How much of it is actually a solid USP that we could put in a few sentences or a paragraph that's used today, and how much is a promise that might happen in the next few years (but might not).

[pchampin]

A very good (video) explainer is actually available here: https://videolectures.net/iswc2012_berners_lee_semantic_web/?t=70 Yes, 2012! And yet it hasn't aged much, IMO.

[melvincarvalho]

Great video. That's around about the time we first started the read write web community group (which incidentally I chaired for 12 years) but that is sadly shutting down now.

I think the good thing about solid is that it leads the way to a web operating system. However that's not an easy thing to explain to people less familiar with the project.

[pchampin]

This was solved by #69 (more specifically this change).