- From: Tomasz Pluskiewicz <tomasz@t-code.pl>
- Date: Tue, 21 Jan 2020 11:03:18 +0100
- To: "public-rdfjs@w3.org" <public-rdfjs@w3.org>
Forwarding from Hydra CG mailing list (With corrections) On 21 January 2020 at 11:00:29, Tomasz Pluskiewicz (tomasz@t-code.pl) wrote: > Hello Hydra, > > Unfortunately conflicting schedule prevents me from joining the con call this time > again. > > However, I would like to share exciting news from my progress of refactoring Alcaeus > (Hydra client) to use the DatasetCore [1] as the underlying data model. > > ---- > > The current version does some awkward transformations to create a rich JS object model. > In the new version, which I now alpha-test as the upcoming v1.0, the internal architecture > changes significantly. I completely remove the JSON-LD processing and the underlying > data model becomes RDF/JS dataset [1]. This will cause some important changes to what > is possible and how representations are stored and processed: > > First, each resource (also ApiDocumentations) will reside in a graph in the in-memory > dataset. Previously, resources were completely detached JS objects. Thanks to putting > them all in a shared RDF-native, quad structure, they will inherently become connected > as the Linked Data they are. So for example, fetching individual members of a collection > or its subsequent pages will build-up a larger graph. > > As consequence, all individual representations can be traversed as if they were a single > data structure, even though they are retrieved as individual resources. > > Also, because the implementation simply wraps around the RDF/JS standard, any compliant > tools can be used to directly access that dataset, either to manipulate the data or add > more triples from other sources. > > Finally, the entire dataset can be trivially serialised a quads and restored, for example > when the browser page is reloaded. The entire data state of an application can be condensed > to the RDF model! > > — > > Those changes will enable previously the awkward or impossible, such as straightforward > caching, working with multiple ApiDocumentation or combining vocabularies with resource > representations. > > > Hope you like what you hear. Let me know what you think. I will share more info when I’m done > with the alpha version and a beta is released. > > Have a good day, > Tom > > [1]: https://rdf.js.org/dataset-spec/ > > >
Received on Tuesday, 21 January 2020 10:03:26 UTC