Cart
[Home] [By Thread] [By Date] [Recent Entries]
>There is Milslav Nic's Graphotron at the www.xvon.org/ site. > >It is amazingly terse: unlike the pic-in-xml graph languages >it is a specialist transformatin tool. It lets you have your instance >data in vanilla XML (i.e. not specified in terms of nodes >or arcs); you specify a transformation using XPaths into >arcs and nodes. Thanks Rick, yep, this definitely falls into the category of 'neato'. As it happens, I'm approaching a similar sort of problem from a slightly different viewpoint - source data in whatever (I'm looking at vanilla XML & RDF) transformation data in RDF, output in whatever (notably SVG). Essentially this won't be unlike the Graphotron, but the transformation will be based on underlying semantics rather than XML structure. I intend to add editability as well, which demands some kind of state representation, so I've got an in-memory graph model, and for the sakes of both simplicity and generality the model I'm using is a generalised graph (rather than e.g. using the RDF model). I was looking at using a graph language for serialisation of the 'base' graph model. At present I've got a RGML serialization, but this format is a little too restrictive in itself - using other namespaces will obviously blow things wide open, but I was hoping to find something where what I needed could be expressed in a single namespace, one which didn't include too much guff. As I'm going to have to work out a schema for specifying the transformations/mappings anyway, at present it looks like it'll be easiest to include the graph representation stuff there - mostly subclassed from RGML. Cheers, Danny.
|
