Taking a break from ontology visualization, we investigate an novel technology out of TopQuadrant that gets users one step closer to realizing the power of the Semantic Web.
Last month, Holger’s blog mentioned that TopQuadrant had released a new “visualization language” for their TopBraid Composer Maestro Edition (Holger Knublauch is TopQuadrant’s Vice President of Product Development). This piqued my interest since any permanent embrace of the Semantic Web outside the research community requires tools and technologies that are both abundant and simple to use.
I downloaded the 2.4.2 version of Composer Maestro and immediately was struck by its similarity to Yahoo Pipes and NetKernel. The basic premise is to use the graph view in Composer to drop instances of modules and connect them together in a linear progression of functionality.
Behind the scenes, each module is an OWL class with predicates that usually define required inputs. Using SparqlMotion’s next object property allows the output from one module (instance) to become the input for a new module (instance). By triggering any module, Java code implements the underlying functionality. There are a total of forty-nine modules, mostly dealing with data transformations (to/from Excel, flat files, XML and RDF in a variety of formats and locations), typical procedural operations (iterations, conditionals) and support for user input. I like the use of SPARQL – yes, of course you knew it was coming to that, didn’t you? ASK clauses are the basis for conditions, CONSTRUCT clauses for filtering and SELECT clauses for iterative behavior.
As the instances and connections fall into place, it is easy to visualize the data flow. There is no forced ontology manipulation, messy rule creation or arduous interaction with semantic languages, the latter which is usually the main complaint of Semantic Web critics. Overall, SparqlMotion is a nice start but it isn’t ready for your mother to begin cranking out semantic applications. TopQuadrant has been upfront that SparqlMotion is in an alpha state and there are a lot of modules I’d love to see: a plugin/interface module where I can develop my own transformations, microformat support and more HTML-friendly exports to name a few.
As far as my exploits, I was a little ambitious at first. I tried to import a complex Excel spreadsheet that’s laden with functions, dependences and tabs. Using the LoadExcelCellInstances module, the only input necessary was a path to the spreadsheet. However, even after verifying the path, Composer could only tell me that it failed to execute based on a NullPointer exception. Trying a simpler spreadsheet with only three columns on a single tab elicited the same response. So, I tried a tab-delineated flat text file (LoadRDFFromSpreadsheet) and got the same response. Moving on, I played around with creating a feed based on George W. Bush from DBpedia (loadRDFFromURL), creating a simple ASK query to determine his presidential qualifications (BranchByAsk) and pushing the resulting RDF either to an XML file (CreateXMLFile) or a Sesame repository (WriteToSesameNativeRepository). Not exactly rocket science, but enough to get a feel toward usability. Serious debugging is limited to running modules individually or serializing results between modules to validate the output. This was somewhat frustrating but hardly impossible.
SparqlMotion is in its early development and I have no doubt it will continue to improve – I see it constantly in Composer. Users will still need to be proficient in SPARQL to utilize certain SparqlMotion modules and adding cut and paste functionality to the form view would be a nice time saver. I was happy to see useful help pages, especially since SparqlMotion is so new. Currently, SparqlMotion execution is all accomplished directly through Composer. Hopefully the ability to run these scripts in a batch mode outside of Composer will enable backend processing and automated services, something that the TopBraid Live server purportedly supports.
There is great potential in SparqlMotion and I continue to be impressed with the innovation that comes out of TopQuadrant. Visit http://sparqlmotion.org/ for more information and even take advantage of a 30-day free evaluation of Maestro. If you’d like to see a six-month re-evaluation, drop me a line and let me know.
- A Tale of Agile Development... of a Standard
- Introduction to: Open World Assumption vs Closed World Assumption
- Summary of 11th International Semantic Web Conference
- Elephant Hunting for Pygmies: How To Introduce Big Data and Linked Data to Your Company