Not to go too deeply into this on jena-users list, but ... I don't doubt that RDF/XML+XSLT is a useful approach in some situations, to meet some information presentation needs.
My experience was in Product Lifecycle Management (PLM), mainly dealing with part master and product structure data, but extending into shop floor (planning), configuration control, and supply chain aspects. There are many different, useful, presentations that can be made from this sort of dataset. Some presentations are easily supported by csv or other record-oriented extracts from RDF dataset. Others require on-the-fly conversion of subgraph extracts. I came to realize that both types of presentations (for web browser delivery) could be handled by relatively light-weight client-side conversions using d3.js. The only drawback is that too much business- and schema-specific selection and transformation logic had to be written in javascript. I would have preferred to use a standard declarative language (like XSLT, only for general graphs) to encode these rules. I'm pretty sure my experience in PLM applies in other domains. The abstract problem is the same: selecting things of interest from an RDF dataset, and applying transformations to display those things with desired presentational semantics (which includes behavioral dimensions). Regards, --Paul On Mon, 2020-03-09 at 08:31 +0100, Martynas Jusevičius wrote: > Paul, > > as I've shown, XSLT is perfectly usable. Admittedly easier on a flat, > non-nested output. > RDF/XML is significant as a bridge format to the XML stack. > > Do you have an example of "useful schema-specific information presentations"? > > > Martynas > > On Mon, Mar 9, 2020 at 3:56 AM Paul Tyson <phty...@sbcglobal.net> wrote: > > > > On Sun, 2020-03-08 at 12:06 +0530, Diptendu Dutta wrote: > > > I have used Jena to generate RDF/XML of the model: > > > > > > <rdf:RDF > > > xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"; > > > xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#";> > > > <rdf:Description> > > > <rdf:subject rdf:parseType="Resource"> > > > <rdfs:label>TheConsignor</rdfs:label> > > > </rdf:subject> > > > <rdf:predicate > > > rdf:resource="http://www.lke.com/lke.owl#shall%20deliver > > > "/> > > > <rdf:object rdf:parseType="Resource"> > > > <rdfs:label>theConsignedUnits</rdfs:label> > > > </rdf:object> > > > </rdf:Description> > > > <rdf:Description> > > > <rdf:subject rdf:parseType="Resource"> > > > <rdfs:label>rulesofinterpretation</rdfs:label> > > > </rdf:subject> > > > <rdf:predicate rdf:resource="http://www.lke.com/lke.owl#apply%20in"/> > > > <rdf:object rdf:parseType="Resource"> > > > <rdfs:label>thisConsignmentAgreement</rdfs:label> > > > </rdf:object> > > > </rdf:Description> > > > . > > > . > > > </rdf:RDF> > > > > > > Do I need to use some other format apart from RDF/XML so that > > > the output my be be visualised in a browser? > > > > > > I looked at "cytoscape.js", Owl2Vowl, and some others. They all require > > > the data to be in some specific format. > > > > > > Are there libraries available for transforming the RDF/XML to a format > > > suitable for display in browser? > > > > > > Which approach would you suggest? > > > > I have not worked in this area since 2015, but it remains the biggest > > problem (and opportunity) in semantic technology. > > > > Must you work with RDF/XML format? Jena provides many other options, > > including csv, json-ld, and sparql results format (srx). These are all > > easier to transform for presentation than RDF/XML. Will you be > > transforming the RDF on the server or in the browser? > > > > Next, do you want to create a static information display, or dynamic > > display that responds to user input? Do you want a primarily text-based > > layout (paragraphs, lists, tables), or graphical (boxes or bubbles and > > lines)? > > > > The approach I found most promising was to use d3.js [1] in the browser. > > This allows you to create either text or graphical layouts, static or > > dynamic. You can make ajax sparql queries, either CONSTRUCT or SELECT, > > and transform the results to HTML5 (including SVG and canvas). What is > > lacking, however, is a standard way to select and transform significant > > graph patterns, analogous to XSLT templates. Perhaps the graph shape > > languages, SHACL and ShEx, could be of some help in this area, but I > > have not kept up in the last 5 years. > > > > One of the big disappointments of semantic technologies is that we > > haven't gotten far past the "connected bubbles" visualization of RDF > > graphs. That is easy to do and most unhelpful. It is much harder to > > create useful schema-specific information presentations from semantic > > data that meet the real needs of the information consumers. > > > > Regards, > > --Paul > > > > [1] https://d3js.org/ > > > > > > > > Regards, > > > > > > Diptendu Dutta > > > > > > Regards, > > > > > > Diptendu Dutta > > > >
