Hi Andy,
> Le 4 sept. 2015 à 10:57, Andy Seaborne <[email protected]> a écrit :
> (snip)
> A good test to add is the null test to get the basic overheads. For short
> queries, more cost goes in HTTP than in the query execution and result
> processing. DESCRIBE{} and SELECT{}.
the first “test” I made was simply looking at the log of fuseki (started as a
sever), and comparing with an output from my servlet (so, comparison on the
server side, wo HTTP cost). But it was very “manual”, and I also realized that
this was inclusing a comparison between tomcat and jetty. That’s why I resolved
to do my tests on the client side, with my servlet and fuseki loaded in the
same tomcat instance
>
> And also for common ground, you can run Fuseki inside Tomcat.
>
> The default RDF/XML output in Fuseki is the pretty printed form
> (RDF/XML-ABBREV).
I had forgotten to say that I set the requested mime type to turtle for rdf
queries, and json for select ones. So, this is not the question here but it is
true that RDF/XML-ABBREV is much longer than RDF/XML. It would be a good idea
to not use it by default, or only for small models.
I’m do think that the probIem is with me, not with fuseki, but I don’t see
where I’m wrong at the moment
Best,
fps
>
> Andy
>
