- From: Konrad Höffner <konrad.hoeffner@uni-leipzig.de>
- Date: Mon, 1 Jul 2024 13:09:47 +0200
- To: public-r2c2@w3.org
Hi, I am a medical computer science researcher at Leipzig University in Germany, at the Institute for Medical Informatics, Statistics and Epidemiology (IMISE) [1]. I have been working with Semantic Web, Linked Open Data and ontologies since 2012 when I did a PhD in Knowledge Graph Question Answering at the research group Agile Knowledge Engineering and Semantic Web (AKSW, also Leipzig University) [2]. I started learning Rust in 2022 and created the RickView SPARQL-less in-memory RDF browser [3] to deploy multiple knowledge graph browsers on memory-starved servers. The aim of RickView is to increase the proportion of KGs that are published according to Linked Open Data standards and to decrease downtimes and discontinuations of KG deployments after the funding period of the associated research project runs out. Soon after, I was asked to support large KGs that don't fit in memory in uncompressed form, which led me to finish the Header Dictionary Triples (HDT) library [4] started by Tim Baccaert and add a Sophia adapter to it, which included profiling and also benchmarking using a fork [5] of the Sophia benchmark suite. The Sophia library allows RickView to query in-memory KGs without a SPARQL endpoint and also to support both compressed (using HDT) and uncompressed in-memory KGs. However, Rust presents some unique challenges, so I am looking forward to a collaborative effort to determine the best compromise for an RDF query API between (1) ease of use, (2) performance and (3) generality of use cases. Konrad Höffner [1] https://www.imise.uni-leipzig.de/en [2] https://aksw.org/ [3] https://crates.io/crates/rickview [4] https://crates.io/crates/hdt [5] https://github.com/KonradHoeffner/rdf_benchmark
Received on Monday, 1 July 2024 11:10:23 UTC