Re: Do you use the W3C RDF validator ?

Hi Pierre-Antoine,

I am not using the RDF validator personally, I use
https://www.easyrdf.org/converter or the JSON-LD playground depending on
the type of RDF I need to check

No one is using the "Validator" because no one needs a validator, what most
people need is a Converter, which also acts as validator at the same time.
And this is also critically missing from the web semantic tools stack.

There are a few unofficial services, but none are perfect
EasyRDF lacks of TriG syntax... Last time I wanted to quickly convert
JSON-LD to TriG. I needed to use JSON-LD to Nquads on the JSON-LD
playground, then Nquads to Trig on EasyRDF...

I did not feel like I was using a mature technology that is 20 years old to
be honest. Especially when the whole point of this technology is to bring
us "interoperability"

The semantic web tooling really needs to be updated from time to time if
the field wants to be taken seriously by other developers. When arriving on
the page it is clear that it is not maintained, and any reasonable person
will wonder "Is it a serious project? Looks like it was abandoned in
1998... I probably should stay away from this"

It does not give any idea of what RDF can be. It makes it looks like there
is only one format for RDF: XML (which is the oldest format, less common
nowadays, people getting into it nowadays usually pick turtle for
readability/compactness, or JSON-LD for web compatibility, or
ntriples/nquads because you can stream them, but I really don't see why
anyone new would want to use XML)

The "reference validator" should support more than the oldest format. Tools
are important for a technology adoption...

From my point of view:
- A minimal modern validator should at least enable users to provide RDF in
various format (the classic xml, turtle, trig, n3, nquads, ntriples,
ideally it could even support JSON-LD).
- And it should enable users to convert from any formats to any other
format (if you can do parsing, you can also do serializing, so why not
doing both?)
- Ideally it should be implemented to work fully on the client (because
decentralization, and scalability, and we have now good JS/wasm options to
parse in the browser now), so that it can be deployed to any CDN without
the cost of hosting a server. If people needs an API we can find a way to
setup a client-side API (look like an API, query like an API, but execution
on the client)


Best,

Vincent


Le mer. 22 nov. 2023 à 13:46, <hans.teijgeler@quicknet.nl> a écrit :

> Hi Pierre-Antoine,
>
>
>
> I use the IDLab Turtle Validator <http://ttl.summerofcode.be/>
>
>
>
> Regards, Hans
>