Posts tagged with: talis

The latest ARC release comes with a couple of non-trivial (but also not necessarily obvious) changes. The most significant (as it involves ARC's resource indexes) is the alignment with the structures used by the Talis platform. ARC's parser output and PHP or JSON formats are now directly processable by Talis' platform tools. The documentation has been updated already, you may have to adjust your code (basically just "s/val/value/" and "s/dt/datatype/") in a few places.

The second major addition is a Remote Store component (documentation still to come) that is inspired and based on Morten Frederiksen's great RemoteEndpointPlugin. The Remote Store works like Morten's Plugin, but supports SPARQL+' LOAD, INSERT, and DELETE (i.e. write/POST) operations.

The third addition is also the reason why the Remote Store (which can be used as a SPARQL Endpoint Proxy) became a core component. I've worked on a draft for a SPARQL-based scripting language during the last months, and the latest ARC revision includes an early SPARQLScript parser and a SPARQLScript processor that can run a set of routines against remote SPARQL endpoints. What's still missing before this stuff becomes more usable (apart from documentation ;) is output templating and some other essential features such as loops. I do have an early prototype running in a local SPARQLBot version, but I probably won't have it online in time for tomorrow's Semantic Scripting Workshop (that I'll try to attend remotely at least). This is really powerful (and fun) stuff that will be available soon-ish. Can't wait to replace my hard-coded inferencer with a set of easily pluggable SPARQLScript procedures.

Other tweaks and changes include a very early hCalendar extractor and a couple of bug fixes that were reported by (among others) the SMOB project maintainers.

As usual, thanks to all who sent in patches, bug reports, feature requests, and stress-tested ARC. I think we're pretty close to a release candidate now :-)

If there was a "most productive SemWeb coder" category in Danny's "This Week's Semantic Web", this week's turn would probably be Keith Alexander's. Last week, he provided no fewer than three ARC2 Plugins:

• An RDFa Serializer
• Utilities for working with ARC structures (e.g. filter, merge, or diff)
• A SPARQL (Re-)Serializer (very handy for checking/adjusting SPARQL queries before they are passed to the store)

While at it, he also implemented a SPARQL+ wrapper for Talis Platform stores.

I think I blogged about Morten's RemoteEndpoint plugin a while back (this one should really become part of the core codebase), but did I mention Peter Krantz' File System Synchronizer? It keeps an RDF Store in sync with a file system directory which enables a really nice option to implement larger RDF editing systems on top of ARC: By using editing tools that work with small RDF files (quick response times and everything) and his plugin, it becomes possible to provide rich query functionality over the whole dataset without the store getting in the way of the publishing tools. RDF index rebuilding can be slow, de-coupling read from write operations and introducing an asynchronous update process is a nice solution.

Awesome stuff.