commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "John D. Ament"<johndam...@apache.org>
Subject Re: [DISCUSS] Commons RDF as new component?
Date Tue, 22 Nov 2016 02:47:18 GMT
Sometimes, we have to believe in github's data.

https://github.com/apache/incubator-commonsrdf/graphs/contributors

It indicates its primarily Stian, a bit of Sergio.  Andy has helped a bit, but I do believe
Peter has stepped away from the project based on the mail archives.  

John

On 2016-11-17 12:11 (-0500), Gary Gregory <garydgregory@gmail.com> wrote: 
> I am more concerned about the size of the community here. Is Paul still
> involved in the code, still wants to be involved, or do we have an SME pool
> size of 1?
> 
> Gary
> 
> On Nov 16, 2016 11:09 PM, "Benedikt Ritter" <britter@apache.org> wrote:
> 
> > I think we can move RDF to Commons regardless of the maturity of the API.
> > We have other proper components which never had a release. RDF already had
> > three.
> >
> > Benedikt
> >
> > Stian Soiland-Reyes <stain@apache.org> schrieb am Mi. 16. Nov. 2016 um
> > 23:52:
> >
> > > Another is that ServiceLoader don't work well under OSGi without careful
> > > hand holding, so it may be best to leave that to the consumer code rather
> > > than say have a static method in the API with a fixed ServiceLoader.
> > >
> > > (I always have trouble with the ServiceLoader instances built in to Java,
> > > which become very sensitive to which thread initates first)
> > >
> > > It might be worth adding the OSGi service @annotations (or corresponding
> > > XML files) so RDF implementations can be used as OSGI services; all
> > Commons
> > > RDF modules should already be bundles.
> > >
> > > OSGi is probably a good fit with Commons RDF as a library can export only
> > > use of interfaces org.apache.commons.rdf.api, but depend on say RDF4J
> > > internally or get any RDF implementation injected as an OSGi service;
> > which
> > > would not be visible to clients of said library.
> > >
> > > And as the Commons RDF API is interoperable, another part of the
> > > application may use Jena (without knowing it) and pass say a Graph to
> > said
> > > library without any problems.
> > >
> > > BTW, anchor link didn't work because GitHub and Maven have different
> > > Markdown flavours, link is instead:
> > >
> > >
> > > https://commonsrdf.incubator.apache.org/userguide.html#
> > Finding_an_RDF_implementation
> > >
> > > As noted in there, care must be taken with reuse of an RDF instance and
> > > labeled blank nodes: there is no RDFFactoryFactory.
> > > rdf.getClass().newInstance() should work for anything from a
> > ServiceLoader
> > > (but not an OSGi Service which uses proxy classes).
> > >
> > > On 16 Nov 2016 10:02 pm, "Peter Ansell" <ansell.peter@gmail.com> wrote:
> > >
> > > > Hi Gary,
> > > >
> > > > There is support for java.util.ServiceLoader with the relevant
> > > > META-INF/services files and default constructors, detailed in the
> > > > userguide.md file currently, but could also be in README.md:
> > > >
> > > > https://github.com/apache/incubator-commonsrdf/blob/
> > > > master/src/site/markdown/userguide.md#Finding_an_RDF_implementation
> > > >
> > > > However, users may be aware of which implementation they prefer and
> > > > use it specifically, to match up with their database drivers or other
> > > > libraries so the direct constructor patterns will be used regularly
> > > > also.
> > > >
> > > > The main reason for that would be that if you use the ServiceLoader
> > > > approach (just using the RDF interface) you don't have object level
> > > > access to the mutators to create native Jena/RDF4J/etc. objects out of
> > > > the Commons RDF objects. There should still be static methods for
> > > > those cases, but they are less simple to use than the mutators in
> > > > general.
> > > >
> > > > Cheers,
> > > >
> > > > Peter
> > > >
> > > > On 17 November 2016 at 02:49, Gary Gregory <garydgregory@gmail.com>
> > > wrote:
> > > > > In the read me I see a link to an example that has
> > > > >
> > > > > public RDF createFactory() {
> > > > >         return new SimpleRDF();
> > > > >     }
> > > > >
> > > > > Is/should a service loader be used to allow pluging in/discovering
> > > > > implementations?
> > > > >
> > > > > Gary
> > > > >
> > > > > On Nov 16, 2016 5:22 AM, "Stian Soiland-Reyes" <stain@apache.org>
> > > wrote:
> > > > >
> > > > > Commons RDF (incubating) has been in the Apache Incubator since 2015,
> > > > > with an aim to graduate into the Apache Commons PMC.   I think now
> > the
> > > > > time has come to propose such a move.
> > > > >
> > > > >
> > > > > This route was agreed with the Commons PMC at the time, and we have
> > > > > had good guidance from our mentors, including Commons PMC folks Gary
> > > > > Gregory, Benedikt Ritter and lately myself :)
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > As noted in the 0.3.0 release today, Commons RDF aims to provide
a
> > > > > common Java API for RDF 1.1 graphs and datasets.  Commons RDF 0.3.0
> > > > > adds API bindings for Apache Jena, Eclipse RDF4J, JSON-LD Java as
> > well
> > > > > as a standalone implementation.
> > > > >
> > > > > http://commonsrdf.incubator.apache.org/
> > > > >
> > > > >
> > > > > Now I think Commons RDF is stabilized and is ready to graduate to
be
> > a
> > > > > component under Commons PMC where we can do a 1.0.0 release to mark
> > > > > the spot.
> > > > >
> > > > >
> > > > > (Concurrent thread on dev@commonsrdf:
> > > > > https://lists.apache.org/thread.html/760b2d58e344e51550dcc9f6452395
> > > > > efe72e97d3dc06169ef72f09c6@%3Cdev.commonsrdf.apache.org%3E
> > > > > )
> > > > >
> > > > >
> > > > > I must admit that community-wise Commons RDF is not as strong as
when
> > > > > the podling started - forming a uniform API across multiple
> > > > > implementations and use cases did mean many hard technical and
> > > > > semantic discussions which caused some people to withdraw for a
> > while.
> > > > > Several has since returned as indirect contributors (e.g. code
> > > > > review).
> > > > >
> > > > >
> > > > > Now that the Commons RDF API has settled, and given that Commons
has
> > a
> > > > > "any ASF committer have write access" policy and an open development
> > > > > mode, I would like to propose here (and at dev@commonsrdf) that the
> > > > > component graduate to be owned by Apache Commons PMC.
> > > > >
> > > > >
> > > > > Views..?
> > > > >
> > > > >
> > > > > Here's a project maturity report
> > > > > https://github.com/apache/incubator-commonsrdf/blob/
> > master/MATURITY.md
> > > > >
> > > > >
> > > > >
> > > > > Want to crunch some code ? Have a look at
> > > > >
> > > > > https://commonsrdf.incubator.apache.org/download
> > > > > https://github.com/apache/incubator-commonsrdf/
> > > > > https://commonsrdf.incubator.apache.org/apidocs/
> > > > >
> > > > >
> > > > > --
> > > > > Stian Soiland-Reyes
> > > > > http://orcid.org/0000-0001-9842-9718
> > > > >
> > > > > ------------------------------------------------------------
> > ---------
> > > > > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > > > For additional commands, e-mail: dev-help@commons.apache.org
> > > >
> > > > ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > > For additional commands, e-mail: dev-help@commons.apache.org
> > > >
> > > >
> > >
> >
> 

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org


Mime
View raw message