Missing full Compliance to the Linked Data Cloud principles

[kuzeko]

To have the bonsai dataset added to the Linked Data Cloud
We need to satisfy the Linked Data Principles

In particular, we need to satisfy the following (quoting from the website):

• There must be resolvable http:// (or https://) URIs.
• They must resolve, with or without content negotiation, to RDF data in one of the popular RDF formats (RDFa, RDF/XML, Turtle, N-Triples).
• The dataset must contain at least 1000 triples.
• The dataset must be connected via RDF links to a dataset that is already in the diagram. This means, either your dataset must use URIs from the other dataset, or vice versa. We arbitrarily require at least 50 links.
• Access of the entire dataset must be possible via RDF crawling, via an RDF dump, or via a SPARQL endpoint.

The missing issue is then the "resolving" of URIs, I think we need something like an .htaccess file or a script that does the correct redirect to our rdf server or to the correct query at our sparql endpoint.

[tngTUDOR]

@kuzeko can you please elaborate a little bit more on the second item ?
Say for example by providing a specific tool that "resolves". For example:
if I do:

wget http://rdf.bonsai.uno/XX/YY

The result must be:

content-type: rdf-ttl 
XXXX: YYYY
data: ZZZ

[cmutel]

First off, this is an excellent issue report, I only wish I was as thorough.

I don't understand the issue itself - is the problem that rdf.bonsai.uno/flowobject/exiobase3_3_17/#C_ADDC doesn't provide any references to the actual datasets? Or is the problem that rdf.bonsai.uno itself doesn't provide a way to traverse its own directory tree?

BTW, htaccess is basically considered an anti-pattern, at least by some.

[tngTUDOR]

I think I understand better now. With content negotiation it would be something like:

curl -H 'accept: text/turtle' https://rdf.bonsai.uno/flowobject/exiobase3_3_17/#C_ADDC

[tngTUDOR]

The rdf contents are statically served (with nginx). I'll look for a solution that allows for content negotiation.

[tngTUDOR]

Right now, without any modifications of nginx, this is possible:

curl -H 'accept: text/turtle' http://rdf.bonsai.uno/flowobject/exiobase3_3_17/exiobase3_3_17.ttl

but I couldn't get a specific \ # reference:

curl -H 'accept: text/turtle' https://rdf.bonsai.uno/flowobject/exiobase3_3_17/exiobase3_3_17.ttl\#C_ADDC 

[kuzeko]

So the best possible (utopic?) solution would be the following:

curl -H 'accept: text/turtle' https://rdf.bonsai.uno/flowobject/exiobase3_3_17/C_ADDC

Should return the result of the query:

SELECT *
FROM <http://rdf.bonsai.uno/flowobject/exiobase3_3_17/>
WHERE {
  BIND ( <http://rdf.bonsai.uno/flowobject/exiobase3_3_17/C_ADDC> as ?s)
 ?s ?p ?o .
}

[kuzeko]

Also note that there is not # in that specific URI (shall we decide on a convention here?)

[cmutel]

We shouldn't be reinventing the wheel, if we need to expose quasi-database endpoints, maybe our URIs should just point to the database. Otherwise, as this is a standard, there should be a standard software that we can just throw .ttl files at and not have to think too much about it, right?

[kuzeko]

We cannot just point to the database.
As you say, the option for NOT reinventing the wheel is to use some additional application software.
For instance, famous endpoints use Virtuoso instead of Jena.
This provides the full stack, e.g.,
https://bio2rdf.org/clinicaltrials:NCT00050895