Re: FYI: prototyping implementation

[Stian Soiland-Reyes <st...@apache.org>]

Great!  (I didn't spot this until now! :/ )

I think it looks very similar to commons-rdf-simple -- do you think
some of your simplifications (e.g. in Graph.add()) should be applied
to simple ?


As I'm looking at defining the hashcodes for COMMONSRDF-14,
would you share any thoughts on why you multiply by 31 here
instead of just returning the uniqueReference() hashCode as-is?

https://github.com/afs/commonsrdf-container/blob/master/src/main/java/org/apache/somewhere/general/BlankNodeImpl.java#L58



On 16 April 2015 at 18:46, Andy Seaborne <an...@apache.org> wrote:
> I have a implementation (WIP) as a basis for experimentation, with a slant
> towards integration across systems (same implementation/different instance,
> or distributed, or different implementation).
>
> https://github.com/afs/commonsrdf-container
>
>         Andy



-- 
Stian Soiland-Reyes
Apache Taverna (incubating), Apache Commons RDF (incubating)
http://orcid.org/0000-0001-9842-9718

Re: FYI: prototyping implementation

[Andy Seaborne <an...@apache.org>]

On 08/05/15 15:50, Stian Soiland-Reyes wrote:
> Great!  (I didn't spot this until now! :/ )
>
> I think it looks very similar to commons-rdf-simple -- do you think
> some of your simplifications (e.g. in Graph.add()) should be applied
> to simple ?

I was trying to just use the external contracts for the RDF terms.  If 
blank nodes have an accessible global uniqueness of some kind, it is 
even-simpler to store as is.

This is the space I was exploring.  The same-object-out-as-in would help 
for integrating systems and general efficiency (no object churn).

If systemX wants to repmap, it can do it itself.


> As I'm looking at defining the hashcodes for COMMONSRDF-14,
> would you share any thoughts on why you multiply by 31 here
> instead of just returning the uniqueReference() hashCode as-is?
>
> https://github.com/afs/commonsrdf-container/blob/master/src/main/java/org/apache/somewhere/general/BlankNodeImpl.java#L58

No strong reason.

I used Eclipse's "generate hashCode and equals" feature but at time when 
there were multiple fields.  Eclispe munges by using "31*"

But now I get:

@Override
     public int hashCode() {
         final int prime = 31 ;
         int result = 1 ;
         result = prime * result +
               ((uuid == null) ? 0 : uuid.hashCode()) ;
         return result ;
}

I may attempt to parse the identifier and if looks like a UUID string, 
parse that.  Depends on how COMMONSRDF-6 turns out.

	Andy

> On 16 April 2015 at 18:46, Andy Seaborne <an...@apache.org> wrote:
>> I have a implementation (WIP) as a basis for experimentation, with a slant
>> towards integration across systems (same implementation/different instance,
>> or distributed, or different implementation).
>>
>> https://github.com/afs/commonsrdf-container
>>
>>          Andy
>
>
>