Re: Implementation report

Off topic a wee bit, but is there more information on the submission process?

The suggested link: seems to mostly describe what needs to go into the output file, but not where to put the results.

Judging from this thread it looks like a Pull Request on the API repo is the right approach for actually submitting the report file?

If that's the right steps to take, I can send a pull request updating those instructions a bit.



From: Gregg Kellogg <>
Sent: Tuesday, March 31, 2020 3:39 PM
To: Julien Lepiller <>
Cc: <>
Subject: Re: Implementation report

I created PR #441 (, since merged) with your implementation report. Note that there were a number of errors that needed to be correct:

* The base IRI for the manifests is set in the context file, and should be, for example, <> not <>.

* It’s worth doing a syntax check on your generated turtle, there were some errors. An example Assertion could be the following:

[ a earl:Assertion;
  earl:assertedBy <>;
  earl:subject <>;
  earl:test <>;
  earl:result [
    a earl:TestResult;
    earl:outcome earl:passed;
    dc:date "2020-03-30T03:33:09+0200"^^xsd:dateTime ];
  earl:mode earl:automatic ] .

(note that closing “]” after the dateTime).

You can try using the “validate” command at, for one.

* Not that it’s significant, but the format of the timezone part of a dateTime should include a “:”, so  "2020-03-30T03:33:09+02:00"^^xsd:dateTime (see It’s not used in the report, but a validator may/should mark it.

* If you add a doap:programming-language entry to your software description, it will be added to your test-subject description (, which would be useful for people looking for an appropriate implementation.

You can see diffs to your source here ( for future reference if you end up re-generating the report.

Anyway, congratulations for achieving such wide conformance, as you can see at

(Note, we’ll freeze the test suite later this week, and you might find some additional tests).

Gregg Kellogg<>

On Mar 29, 2020, at 7:11 PM, Julien Lepiller <<>> wrote:


attached is an implementation report for guile-jsonld, an
implementation of the JsonLD API in the GNU Guile language.

I haven't yet released a version of that library, so this report is for
the current master branch. If you need a release, then I'll make one
from that commit and give it a proper version number.

I haven't used or produced turtle documents before, but I tried to
produce a document that hopefully makes sense. I'm not sure if what I
did is correct, especially in the results, so please tell me if
anything is wrong, so i can correct it.

My library implements only JsonLD version 1.1, so I skiped tests
designed only for JsonLD 1.0. I still included them with a result of
earl:inapplicable. Is it correct, or should I simply not include
skiped tests?

I haven't implemented any of the features tested by the to RDF and from
RDF test suites, and my implementation of features tested by the HTML
test suite is lacking (my main issue being that there doesn't seem to
be an HTML parser for Guile, I used an xml parser instead, but it's not
enough). I haven't included any results from these three test suites.

Thanks in advance,
Julien Lepiller

Received on Tuesday, 31 March 2020 19:54:36 UTC