Re: tracking state changes in a temporal read-write web

> > I have a suggestion for a simple decentralized use case:
> >
> >     There are 2 agents running instances of the same application,
> > where the instances are peers since the application includes both a
> > server (with an RDF storage backend) and a client and can communicate
> > both ways.
> >     One of the agents accesses (dereferences) an RDF document on the
> > peer application, and stores that data in its own application.
> >
> > And that's it, to begin with. The intention is that now the agent can
> > cross-reference the new data with the rest of the data in its
> > application, e.g. using SPARQL if the storage supports it.
> > Authentication, authorization are of course also in this picture, but
> > they are orthogonal, so for the sake of simplicity we can skip them
> > for now.
> >
> > Is that too simplistic? Then please show me an RDF-based app that can
> > do this out-of-the-box.
>
>
> Nice and simple use-case.

I'm glad we're in agreement here, that doesn't happen so often ;)

>
> Here's a suggestion re most basic RWW use-case:
>
> A simple Application deployed in Single Page Application mode (ie.,
> HTML, CSS, and JS) that can achieve the following:
>
> 1. Authenticate using a variety of protocols
>
> 2. Insert, Update, or Delete Data using a simple data entry form or via
> SPARQL; authentication is multi-protocol thereby offering choices;
> storage options included a File System or DBMS; and all this subject to
> ACLs in place.
>
>
> Example:
>
> [1] https://github.com/OpenLinkSoftware/single-page-apps

I checked the demo here:
https://openlinksoftware.github.io/single-page-apps/data-entry-form.html

Technically this example might match the description in my use case.
As does Atomic Data (https://atomicdata.dev/) mentioned by Jonas.
But my take is that these are demos and nowhere near consumer
products. There is *a lot* lacking, especially in terms of UX -- it
doesn't seem like it's evolved much since the beginning of Linked Data
10+ years ago.
First of all, we heard a number of times from commercial users that
they are not interested in seeing the technical details of RDF. So
editing raw triples (rather than entities with properties), displaying
raw URIs -- that's a no go from the start.

Now take a look at Roam Research [1], Notion [2] and their UX. Heck,
even at the Freebase demo from 2008 that shows parallax navigation
[3].
The question is: how do we close the UX gap to the level of those
products, while building with the basic read-write Linked Data (RDF
CRUD) and SPARQL building blocks that we have at hand?

Hint: we have some opinionated ideas [4].

[1] https://roamresearch.com/
[2] https://www.notion.so/
[3] https://vimeo.com/1513562
[4] https://atomgraph.com/blog/

Received on Friday, 21 May 2021 18:30:05 UTC