[Hit] Multiple DAO

[Kyle Braak [GBIF]]

Dear Julien, 

The DAOs with .synchronise deal with synchronization (synchronizing raw occurrence records). The other set of DAOs deal with extraction (extracting raw occurrence records into occurrence records). 

Some background. Before the HIT, indexing at GBIF was done using this code. When developing the HIT, the synchronization part was rewritten, and the extraction part remained the same. To not risk breaking extraction, it was safer to just wire up the old extraction DAOs unchanged, and then code a new set of DAOs for synchronization. 

After a couple years, the extraction part was rewritten anyways, done independent of the HIT using Hadoop. Therefore, the synchroniser-gbif was branched into synchroniser-gbif-refactor, with all the extraction code completely removed. 

Another thing you will likely notice is that the HIT can no longer communicate with the GBIF Registry. 

Over the course of many years, the HIT has constantly had to adapt with the evolving GBIF Registry. First, it talked with the original GBIF (UDDI-based) registry, then with the GBIF GBRDS Registry, and lastly with the GBIF (dataset-aware) Registry. 

This month GBIF has discontinued its reliance on the HIT for indexing, relying instead on a new suite of tools and technologies such as HBase that will enable real-time indexing. This new indexing communicates with the latest version of the Registry: the GBIF (real-time) Registry. 

Therefore, the HIT is now out of date. Furthermore, as mentioned above, the latest branch of the synchronizer project doesn't include any of the extraction code. As you are aware, extraction is needed to be able to populate the database serving an adapted GBIF Data Portal. 

The HIT will always remain a free and open-source project. Therefore, if you are interested in updating it to use the latest Registry, and adding the extraction code back in, I could try to guide you. There are a handful of existing HIT installations around the world that would also benefit from your volunteer efforts to update the project.

With kind regards,

Kyle Braak
Developer
Secretariat of the Global Biodiversity Information Facility (GBIF)
Universitetsparken 15, DK-2100 Copenhagen Ø
Denmark, Europe
+45 353 21479 (work)
Skype: kyle.braak
www.gbif.org

On Oct 31, 2013, at 1:18 PM, Julien Husson wrote:

> hi guys,
> 
> I'm working on a custom synchronizer for my own portal with an Oracle Server.
> 
> I don't understand why DAO are duplicate for example :
> 
> org.gbif.harvest.portal.synchronise.dao.ImageRecordDAO
> org.gbif.portal.dao.ImageRecordDAO
> 
> it's the same thing for DAO Impl and Model
> 
> Is there a cause ? What's the motive of this ?
> 
> In advance Thx you,
> 
> Julien H.
> _______________________________________________
> Hit mailing list
> Hit at lists.gbif.org
> http://lists.gbif.org/mailman/listinfo/hit

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.gbif.org/pipermail/hit/attachments/20131107/12a04a03/attachment.html