On 17/12/14 18:34, Reto Gmür wrote:
Hi Andy
I'm aware of this code it doesn't imho satisfy the clerezza stated goal
(since the incubation proposal) to provide "An API modeling the W3C RDF
standard without any vendor specific additions".
Many systems make that claim.
I don't know what vendor specific additions might be.
The issues discussions around "why?" for commons-rdf are very useful
here. What matters what a new API can do for users. commons-rdf aims
to make base storage switchable so that you can have portable algorithms
or mixed systems.
Jena's goal is :
[[
the creation and maintenance of
open-source software related to accessing, storing, querying,
publishing and reasoning with semantic web data while
adhering to relevant W3C and community standards.
]]
Clerezza has it's own take on the specs, especially around blank nodes,
leaning and default graphs having names.
All communities have experiences that have taken them in different
directions.
Yet in the current version
the naming has become closer to the wording in the spec so we could
partially use the results of this project.
Yes - I was on the WG.
As concrete shortcoming of github-commons I see:
- It adds things that are not in the API and make it less general purpose,
like adding blank node identifiers from concrete syntaxes
Untrue. The javadoc explains this
"This is not a serialization/syntax label."
- It doesn't define the identity criteria (which quite clearly should be
different for mutable and immutable graphs) for the defined types
Yes, it does. Term equality.
If you want term equality for graphs how do you solve the size problem?
>100e6 triples in a graph.
As you know, Jena has put a lot of work into bnode isomorphism
efficiency and even then makes no pretence of graph equality based on
term or value equality.
Andy
- It doesn't extend the java collection APIs
I've created https://issues.apache.org/jira/browse/COMMONSSITE-80 and I
will commit a proposal created from the work at clerezza and github-commons.
Cheers,
Reto
On Wed, Dec 17, 2014 at 4:10 PM, Andy Seaborne <[email protected]> wrote:
The system you referenced is now at:
https://github.com/commons-rdf/commons-rdf
and I hope you'll consider it for Clerezza.
Andy
On 17/12/14 15:23, Reto Gmür wrote:
Hi,
This sounds like a great occasion to publish a revised version of the core
RDF libraries in clerezza as Commons RDF.
This has been the topic of several discussions already. The Clerezza RDF
libraries are strictly based on the RDF concepts and does not introduce
auxiliary concepts (like BNode labels) that narrow the field in which the
library is useful. Compared with the APIs provided by triple stores the
Clerezza API can thus more broadly be used for use cases benefiting from
the RDF model.
I see two major TODOs:
- Renaming: The current naming puts great emphasis on technical
correctness
at the expenses of matching the colloquial use of the terms. The API
should
be simplified to have Graphs and ImmutableGraphs rather than
TripleCollections, Graphs and MGraphs [1].
- RDF 1.1 Adaptation: the identity criteria must be redefined and probably
the class structure adapted for the identity of no language plain literals
and xsd-string typed literals
Cheers,
Reto
1.
http://mail-archives.apache.org/mod_mbox/clerezza-dev/
201406.mbox/%3CCALvhUEWd_qYqLAANS6oeVVnR1ndzEv9FSSrO0s6
[email protected]%3E
---------- Forwarded message ----------
From: Benedikt Ritter <[email protected]>
Date: Mon, Dec 15, 2014 at 4:05 PM
Subject: [ANNOUNCEMENT] Apache Commons grants write access to all ASF
committers
To: [email protected], Commons Developers List <
[email protected]>
Dear fellow committers,
The Apache Commons Team is pleased to announce that write access to the
Apache Commons Subversion and Git repositories has been granted to all ASF
committers.
Apache Commons is an Apache project focused on all aspects of reusable
Java
components. As such, the components maintained by the Apache Commons
project
may be of interest to a variety of other Apache projects.
The Apache Commons community would like to invite you to share and
maintain
useful code.
While Apache Commons is a Commit-Then-Review community, we would consider
it polite and helpful for contributors to announce their intentions and
plans on the dev mailing list [1] before committing code.
Have fun,
Benedikt Ritter,
on behalf of the Apache Commons Community
[1] http://commons.apache.org/mail-lists.html
--
http://people.apache.org/~britter/
http://www.systemoutprintln.de/
http://twitter.com/BenediktRitter
http://github.com/britter
