- From: Ruben Verborgh <ruben.verborgh@ugent.be>
- Date: Wed, 26 Nov 2014 08:16:20 +0100
- To: Kjetil Kjernsmo <kjetil@kjernsmo.net>
- Cc: public-hydra@w3.org
Hi Kjetil, > I was initially skeptical to the whole LDF idea, because it felt too much like the early federation implementations that had to do thousands or tens of thousands of HTTP requests to answer trivial questions, but clearly, the requirement that you need to expose cardinality helps a lot. Exactly! And we strive to make the trivial questions really fast, because simple questions are the only thing clients can ask (so nobody else can ask more difficult questions either, reducing server cost). Furthermore, the current exercise is: how far can we get if the server only offers paged triple pattern data and counts? And while it's interesting to see how far we can get, this is not the final answer yet—and there never will be one. A next question could be: if we add that and that metadata, how much faster does querying become? How difficult does it become for the server? How does federation improve? In other words, there's a whole axis still to explore; yet it's interesting to see how much we can already do with so little :-) Best, Ruben
Received on Wednesday, 26 November 2014 07:16:49 UTC