| Link from here | Target |
|---|
| swrc-ext:authorList |
| rdf:type | swc:Poster |
| foaf:maker | person:alan-ruttenberg |
| swrc:url | http://www.webont.org/owled/2007/PapersPDF/submission_46.pdf |
| swc:uuid | 85514414-6397-4ac5-a3a8-7ba4f43ebdf6 |
| swrc:keywords | Active values |
| swrc:keywords | Computed properties |
| swrc:keywords | Domain specific languages |
| dc:title | Hey! You Got Imperative in My Declarative, or A Mashup Made in Heaven: Making OWL
friendlier with Javascript. |
| rdfs:label | Hey! You Got Imperative in My Declarative, or A Mashup Made in Heaven: Making OWL
friendlier with Javascript. |
| swrc:keywords | Javascript |
| swrc:abstract | We propose that OWL 1.1 incorporate the use of Javascript. By choosing to make use
of Javascript within OWL, adoption and utility of OWL might be improved in two different
ways. First, it could be used in a facility that lets users specify Javascript code to
translate domain specific languages into OWL. Second, Javascript is well known to a large
number of web developers. By enabling OWL documents to have property values that are
computed by Javascript functions, we extend the language in a useful way, and encourage the
use of OWL in different applications than it might otherwise be used. OWL needs a method of
extending its syntax to enable concise expression of domain statements without compromising
its expressiveness. For example, a current debate over the whether to use OWL as the
"native" format for the OBO ontologies, is driven (away from using OWL) by
OWL's unappealing syntax and the relative ease of understanding OBO's
current format. We describe a model for how to use Javascript, included as part of an OWL
ontology, to parse domain specific languages into native OWL. Two levels of operation are
distinguished, the first a lexical translation into a language which extends the functional
syntax, and a second macro expansion which translates the extended functional syntax to
expressions that only use the defined OWL vocabulary. For an implementation of computed
property values, we propose that they have a status similar to annotation properties, in
that they are not reasoned over. Instead they can be computed in terms of non-annotation
property values. In this way, we allow for useful expressivity gains without complicating
the OWL reasoning or imposing a complicated evaluation model. We discuss what access to the
environment, such as the ability to query against the ontology, could profitably (and
safely) be had by these scripts. Use cases motivated from experience in validating BioPAX
and from other applications are provided. Javascript interpreters are available for use from
within the programming languages that the major reasoners are implemented in, and should
therefore not pose an excessive burden on reasoner developers. |
Explore this resource elsewhere:
