– DRAFT –
SPARQL TF

Scribe?

Process and Admin

AndyS: nothing for Process and admin from my side
… general idea for today to look at the tests and then to look at the PR

PR #177

<gb> Issue 177 not found

james: I sent an email regarding the tests
… about where to put the results
… haven't had much time to look at the results
… Question is whether it makes sense to add a folder for putting the results
… for different query engines
… to collect the results all in a document
… in order to be able to compare results across implementations

AndyS: I don't think it is a good idea
… we already have a format for the manifests
… engine implementers have already invested time supporting that
… I think you are asking for reports, not results

james: I have not used W3C machinery recently
… but I don't think there is a facility to compare

AndyS: results are expected results not reporting results
… for reporting, there is the EARL format

james: but that's for a single implementation only

AndyS: yes, but these can be aggregated

james: I want the ability to view them side by side

<AndyS> https://www.w3.org/2009/sparql/implementations/

pfps: I EARL can be used to say what your system is, then this should be fine
… but this doesn't have EARL data in it.

AndyS: I believe there is code that takes EARL reports and creates HTML representation.

james: Another thing that would be useful for the discussion would be to have the query and the result produced by the implementation; EARL has only pass/fail

pfps: Yes, that would be good to have

AndyS: If you could take the action to make this work, that would be great.

james: Yes, I am proposing a directory structure for it. I am implementing something on top of it.

pfps: Better would be to add a new field to EARL to include the produced query results in the EARL reports.

james: My suggestion is a directory tree that parallels the directory tree of the tests, into which implementers put their query results.

pfps: Still better if implementers would extend their EARL output

<Zakim> Tpt, you wanted to comment on over engineering

pfps: I will look into EARL

Tpt: Can't we be lazy?

create a variant of the test per possible output this way we can see which behavior each implementation chose to have

james: pfps, can you send me your results

pfps: no

james: please send them

pfps: For the three that I managed to run, there were lots of variations and divergence from the standard

AndyS: Why don't you create gist or a wiki page for it?

pfps: yes, I can do that

AndyS: We now have five test directories, one for each of the five EXIST-related issues

<AndyS> https://github.com/afs/SPARQL-exists/blob/main/tests/exists-2-def-var/exists-bad-syntax-01.rq

<AndyS> https://github.com/afs/SPARQL-exists/blob/main/tests/exists-2-def-var/exists-good-syntax-01.rq

<AndyS> https://github.com/afs/SPARQL-exists/blob/main/tests/

<pfps> the idea is that you could have several "tests" with the same query but different "expected" results

Tests for issues 1 to 5

<AndyS> https://github.com/afs/SPARQL-exists/blob/main/tests/exists-1-empty-bgp/exists-empty-bgp-02.rq

olaf: Only tests with EXISTS in FILTER?

AndyS: yes, at the moment

olaf: It might make sense to also have tests with FILTER in BIND

Updates tp PR 177

Topic for next time

AndyS: For next time, one item can be pfps' work
… on the tests

pfps: What exactly?

AndyS: Your action on the tests