Drupal Semantics presentation in NYC

Scor's presentation:

Just installing RDF CCK, you get all this for free.

It prefaces every field with site:

It's your own local vocabulary.

you can see the entire site schema vocabulary at /ns#

The source

RDF and RDF/XML are two different things.

On the abstract level, you have RDF, a graph representation of the knowledge.

For machines to

n-triple

turtle is currently the one

RDFa was just standardized, it's a serialization built into the HTML.

It's all the same thing on the RDF level.

So far we have had a site with it's own local vocabulary.

Now it gets more exciting.

Dublin Core, FOAF, SIOC, SKOS, etc.

Sends a Sparql query to a sparql endpoint to get all of the properties of terms of a given vocabulary.

For instance,

When you install EVOC, it will get DC, FOAF, and SIOC automatically.

autocomplete field that will look through all classes

I store the domain and range, but am not restricting RDF property selection by domain and range yet.

Questions:

Vocabulary data.

Marco: My clients always want a UI ontology.

It's coming.

Any inferencing?

No. Was talking about having a hierarchy of content types, and having the parents

Classes are added to content types.
Properties are added to fields.

(With classes Upper case and properties lower case

I plan to have more help, to reuse domain and range, so that the suggested properties only match what is correct to put there.

(The autocomplete acts on both the name and the comment.)

I see a potential use for this in managing some of our profile data-- that is currently stored under a file system. How flexible is this to have sub-nodes-- could you have a node class that describes a single client, and then departments

Scor: No, CCK allows you to share fields

If we could point directly to a subdirectory,

Is the theming system semantically aware?

So far it has nothing to do with the theming.

Is the theming system aware of

For instance, everything that is geocoded would display as a map.

Can you query your own data?

Yes, I show later?

Q: Sparql query that results in a sparql query. Can you have

You can use any vocabulary you want to use for your mapping. And Neologism module allows you to publish your own vocabulary online.

Set of values defined through a Sparql query -- if you knew where to look up the

You could have a constraint for your type, saying that it can only have provinces of Canada.

Scor: It's more than a range restriction, that's a restriction on [what it selects].

CHG: It can be implemented as a Sparql query to an external source.

At the top I have the prefixes definitions:

rdf
site
rdfs
foaf
owl

Any information about the page here? There is a bug, it suggests i was created yesterday.
This URI resolves to me-- if I want to do it completely well I should have the link [to me] as well.

Make it clear that URIs refer to the same person-- owl:sameAs

Q: Can you have an e-mail hash instead of e-mail?
S: On the CCK level, you can have computed fields.

Improved schema. Still an RDFS class-- but also now a subclass of FOAF person.

rdfs:subClassOf <http://xmlns.com/ >

That way our site is still compliant

Great way to give people who don't care how

Have RDF out of the box. If you care about it, you can then do your own mappings. And the idea is for Drupal core to provide intelligent

How to do something with this RDF?

This is a different interface, a visualizer running on Java, written by Andreas Heart, colleague at DERI.

The system knows tons of information about him, all from RDF.

homepage
made (posts on the site)
pastProject

I have a faceted search yere, I can display all the people that I know.
then I can go from here and show all the posts that all the people that i know, wrote.

This is all done through RDF.

Can also view it as a table.

Stephane Corlosquet knows Andreas who wrote this and this

The system has been crawled to get a full RDF import. This interface has not just one site but thousands of sites in it. It browses the data and keeps it.

This tool is not open source?

No! It's still in alpha, it will contain lots of information. He's crawling data from six degrees from Tim Berners-Lee. Currently 12 million statements, almost one million documents.

Neologism - http://drupal.org/project/neologism

Drupal is widely available.

Advantages and disadvantages of allowing people to use a vocabulary that should have a clear schema attached to it, and they may use it incorrectly.

Need for validation as your crawling resources.

At some point, need some reasoning. Unfortunately PHP isn't very good at this, but could use an external service running Java.

Coming out of PHP and relational database -- do you see the need for

Q: Inferencing and pulling in restrictions on values for resources...

Marco: type checking or presentation purposes? Background in database?

Q: Advantages in Python and Django?

Marco: Yes, there are better languages for working with this.
Type checking on the database schema level. Don't have to do it at the RDF level.

Great value of Drupal it makes it more accessible -- the point in the first presentation, I think that was the most valuable point.

Applications solving major economic problems by people being able to mark up

Mixing up creation of data and use of data.

Drupal can already be great for creating data. Other things can use the data. And probably nothing beats Java at the moment for using this data.

How to build in, for this module development, checking for people creating things.

Benjamin Novak, the author of the ARC2 library, speaking in May.

Scor: like broken HTML, there are tons of pages

These are two different things, how to visualize the data

Scor: I'm in touch with the Yahoo Searchmonkey guys

Resolution