We introduce Serverless Data! Datasets that can be used without a developer setting up a dataset. However, the developer would also like any kind of flexibility to provide any kind of feature. Do data publishers then have to set up any kind of possible querying API? I don’t believe so: hypermedia data fragments to the rescue!
Download the full data dump of addresses in Flanders from: https://swarmstoraccountprod.blob.core.windows.net/oslo-ttl-dumps-production/adresregister.tar
Check out the autocompletion with fragments demo at http://126.96.36.199/site/index.html (if this link is broken, inform me: we’re going to give it a new home)
The specification: https://github.com/pietercolpaert/TREE
Full paper with benchmarks: http://pieter.pm/icwe2020-autocompletion/