Hi Benedikt,
On 21/07/14 13:12, Benedikt Ritter wrote:
first of all, welcome the Apache Commons dev mailing list. It's nice to
hear that you have decided to share code with others using the Apache
Commons project.
Consider this contribution our (Andy, Peter and mine) two cents to such
important project. Thanks for having us on board.
In the past we had the policy, that the commons sandbox is free for all ASF
committers. So you could have started coding at commons right from the
beginning ;-)
Since you already bringing a group of people who are willing to work on the
component, I don't have a problem with letting commons-rdf join the sandbox.
Sorry, I tried to find it, but in the web site there is no information
about such procedure. Do we need to start a vote and get a consensus
from the Commons PMC?
Using git may be an issue. We had long discussions about svn vs. git and we
came to the conclusion that using the one or the other should be decided on
a component basis by the developers of the components. So far only Commons
Math is really taking action to migrate to git.
If you really want to use git, we need to figure out how we can integrate
that with our release process. Having you join us, may be a chance for us
to learn from more experienced git users ;-)
Definitely we do need git, otherwise some tasks would require extra time
to fuss with the weird way that svn does things after working with git.
I'm not aware of the long discussions you mention. But of course I'm
willing to help you guys with the migration of those subprojects that
want to try.
I'm not sure about the git PRs. Does that involve github? What about the IP
in such a case? It legal okay with that?
Infra has already worked on providing the basic infrastructure. Then
each project is free to manage it as they prefer. What we did in
Marmotta, for instance, is to force to file issues in Jira for all PRs
(so having the regular IP in place).
About the commons-rdf code itself: I only took a very brief look. As far as
I understand this is only supposed to be an API, hence there are no
implementation classes, right? One thing that caught my eye is
BlankNoreOrIRI... Is this defined as such in the standard?
You are completely right. There has been a long discussion about that
aspect, see issue #9 fir further background:
https://github.com/wikier/commons-rdf/issues/9
Because there is not such official term for what a RDF subject could
have a triple.
It just feels wrong, and I'm seeing people doing:
if(term instanceof BlankNode)
BlankNode node = (BlankNode) term
else
IRI iri = (IRI) term
It is not expected to have classes implementing only that interface, so
it's only useful for the methods signature. That's indeed a problem in
Java, so we may have to figure out better design solutions.
But as I said, I only too a very brief look and I don't know RDF too well.
Further more, where are the license headers in your files?!
Already working on it ;-)
https://github.com/wikier/commons-rdf/issues/28
Thanks for all your comments.
Cheers,
--
Sergio Fernández
Partner Technology Manager
Redlink GmbH
m: +43 660 2747 925
e: sergio.fernan...@redlink.co
w: http://redlink.co
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
