= Bolsena 2012 - draft agenda = For the 5th time, the [http://wiki.osgeo.org/wiki/Bolsena_Code_Sprint_2012 Bolsena OSGeo Hacking Event] is taking place. This page lists some ideas that can be discussed within the GeoNetwork faction. Please add any idea you have here ! [[Image(ConventoBolsena.jpg)]] [[PageOutline(2-3,,inline)]] == On going projects ? == == Community == === Committees === * Review what's improved, for the Advisory Board and the Program Steering Committee, since last year * Many times, PSC members do not vote on proposals within the [http://trac.osgeo.org/geonetwork/wiki/PSC#DetailedProcess allotted time frame] for that. How can we improve on that. === mailing lists === The percentage of mails to the mailing lists with 0 responses (counted from the last max 105 mails) is: * [http://osgeo-org.1560.n6.nabble.com/GeoNetwork-developer-f3874395.html developers list]: 47% * [http://osgeo-org.1560.n6.nabble.com/GeoNetwork-users-f3860293.html users list (english)]: 56% * [http://osgeo-org.1560.n6.nabble.com/GeoNetwork-users-francophone-f3880536.html users list (french)]: 44% * [http://osgeo-org.1560.n6.nabble.com/GeoNetwork-users-espa-ol-f3860136.html users list (spanish)]: 56% It's like [http://en.wikipedia.org/wiki/Coin_flipping flipping a coin], asking something on our mailing lists. On the largest (english users), as on the spanish users, you might as well think you should forget about it as the [http://en.wikipedia.org/wiki/Odds odds] are against you getting even 1 single reply. Is it an idea to establish a Mailing List Governance Committee whose task it is to monitor any mails to all of the lists that go unanswered for more than say 2 days, and then approach people to ask if they would respond? It's great to have so many subscribers but it would be even greater to have a substantially larger number of emails responded. By contrast, as an example, the same count on the [http://apache-wicket.1842946.n4.nabble.com/Users-forum-f1842947.html Wicket users list] shows 18% unanswered emails. == On going proposal == * See [wiki:proposals] page * [wiki:proposals/DCATandRDFServices Data Catalog Vocabulary and RDF services] == Discussion on architecture updates == === Security === * Status on http://trac.osgeo.org/geonetwork/wiki/proposals/ImprovedSecurityArchitecture * CAS support & LDAP improvements (Jesse & Francois) === domain model === Considering GeoNetwork's functionality, you might expect there exist a class called Metadata, representing metadata. Sadly, this is not the case. The lack of a [http://en.wikipedia.org/wiki/Domain_model domain model] makes doing changes to GeoNetwork very much more difficult and encourages source code pasta. Let's model our domain, and while at it, use [http://www.hibernate.org/ Hibernate] to have the domain classes hide the database. === logging === There is no point in maintaining a [http://en.wikipedia.org/wiki/Do_it_yourself home-made] logging facility when there are de facto standard libraries for this. Let's replace GN logging with [http://www.slf4j.org/ slf4j], and back it by the venerable [http://logging.apache.org/log4j/1.2/ log4j]. === i18n === We have [http://en.wikipedia.org/wiki/I18n localisation] in a very large number of xml files, in the database, and in yet different locations for Javascript. It would be most welcome to have 1 single place for all translations. At the very least, is there anything against merging all the xml files into a single file per language? === Schema plugin === * Schematron http://osgeo-org.1560.n6.nabble.com/Schema-plugins-and-schematron-rules-td4962553.html === Lucene indexing === Actually, only fields defined in tokenized section in config-lucene.xml are analyzed by !LuceneQueryBuilder when used in searches. This means for example, if define a new analyzer similar to !KeywordAnalyzer to lowecase terms (in this case the field should not tokenized), !LuceneQueryBuilder will no apply the analyzer to search term, not working well the search if term provided has uppercase letters. Review !LuceneQueryBuilder to apply the analyzers properly, independent if a field is defined in tokenized section in config-lucene.xml or not. == Webspecs testing == The last code spring I (Jesse) created a small suite of integration tests by taking the existing Geonetwork integration tests and combining them with some tests back ported form Geocat integration tests. I also started a maven module that would run the tests when ran with the profile integration-tests. Since then I have updated the tests and fixed bugs based on those tests. I would like to commit this work and back-port more tests. In addition I have started working on some javascript tests. I would like to see them integrated into the build. == Repository migration to !GitHub == Discussion about advantages/disadvantages of repository migration to !GitHub. Tools like [http://subgit.com/index.html/ SubGit] seem can help for smooth migration, keeping in synch SVN-!GitHub repositories. == Wishlist == == R&D == == Presentations == confirmed: * Clustering GeoNetwork, by Jose García & Heikki Doeleman * Pessimistic locking and metadata lifecycle, by Jose García & Heikki Doeleman * Metadata difference viewer, by Jose García & Heikki Doeleman * Monitoring Webapps (Geocat usecase) by Jesse * Git for SVN users by Jesse requested (who wants to do it?) : * [http://hadoop.apache.org/ Hadoop], [http://teddziuba.com/2010/03/i-cant-wait-for-nosql-to-die.html NoSQL] * [http://lucene.apache.org/solr/ SOLR in practice]