All of us know some difficulties with ZEBRA mainly for UNIMARC flavour. Some friends are pushing SOLR as an alternative but now we are seeing ELASTICSEARCH. What you think about it ?

I would prefer solr, since it seems that the community using and supporting it is quite big. The project is under Apache + consider also that vufind is using it. I would prefer a single indexing engine.

I have another question about replacing ZEBRA. How do we maintain Z39.50 server ? best regards,

That is an interesting q.! I think newgen lib is using solr and has a z39 server but only for marc21. Harry

All of us know some difficulties with ZEBRA mainly for UNIMARC flavour. Some friends are pushing SOLR as an alternative but now we are seeing ELASTICSEARCH. What you think about it ? I heard good things about elasticsearch, behind, it is always Lucene

Le 10/03/2013 20:04, Rafael Antonio a écrit : like Solr. The abstract search layer can accept as many search engine as you wish, you would just have to implement the layer. But today, the point is, before doing anything more, the layer should be completly working with zebra. I know Jared is doing developpment around Zebra refactoring.

I have another question about replacing ZEBRA. How do we maintain Z39.50 server ? (We are not talking about replacing Zebra.)

In BibLibre Solr implementation, we developped a Z39.50 server based on Net::Z3950::SimpleServer and Regexp::Grammar.You can have a look here: http://git.biblibre.com/?p=koha;a=blob;f=C4/RPN2Solr/lib/RPN2Solr.pm;h=5d062... Have a good day, Claire.