BibWorm / News: Recent posts

It's been a long time without a news update, sincere apologies for that. Development of the BibWorm Core (current focus on the Data Abstraction Layer) is progressing slowly but surely. Tobyhist currently has little time to work on this project and constant reviews and optimizations slow progress down somewhat but most classes have already been implemented. You are very welcome to review the code and offer suggestions via the project forums. A documentation for the core will follow as soon as possible...

Today, the publication of the class reference for BibWorm's core was begun. You can consult it at http://bibworm.sourceforge.net/development/core_classes.php As this document will serve as the main design guide for BibWorm's API and data abstraction layer, its completion enjoys high priority.

Finally the BibWorm project website is online. It's still under construction, but the development section (http://bibworm.sourceforge.net/development/) already contains a very general outline of the project. The many gaps still there will be updates as soon as possible. In the meantime please do complete our survey at http://bibworm.sourceforge.net/survey/public/survey.php?name=initialsurvey Your information will help us build BibWorm solutions with the features YOU need.

Yesterday, the SVN repository for the BibWorm project was restructured. Old C# code has been moved to the legacy subdirectory of trunk. In addition to this, I uploaded the first few lines of Java code, which form part of BibWorm's core, more precisely the data abstraction layer. Pelase note that this code has not been tested yet due to the many cross-dependencies on other classes yet unwritten.

This is a new tech demo release of BibWorm to illustrate the potential of the plugin system. Functionality itself is in many ways even further reduced than in in previous releases but do play around with the plugins (i.e. move them somewhere else) to see the difference. Further explanation is in readme.txt contained in the ZIP archive.

BibWorm has undergone a major replanning, resulting in a fundamentally different approach. First of all, it will be stromngly plugin driven, which means that essentially BibWorm will consist of an API library also containing the essential data abstraction layer, a loader to search and initialize plugins (including custom user interfaces) and default implementations for a user interface, a settings provider (based on XML configuration files) and probably a basic engine for formatting citations and references based on XML style definition files. This API library will serve as the core for a standalone client as well as a SOAP-based webservice and a web client. I will post a more detailed plan and roadmap soon. Development is currently concentrating on the core library (the abstraction layer and API components). With the change of direction, BibWorm will now migrate to Java to avoid (or at least reduce) problems of portability. My intention, however, is to make a .NET-based core available as well, primarily with the goal of providing full support of Microsoft's Ink for Tablet PCs, but it may also become the favourite implementation for webservices run on IIS. A lot is going on here so watch out for further updates.

Today's release as well as the updated sources in the SVN repository now include the SQLite .NET data provider, which is now pre-registered with the app.conf for the data abstraction layer module. There shouldn't be any need to install it into the Global Assembly Cache and modify machine.conf aynmore. I hope that will make things a little easier. By the way, any feedback, in particular on the basic idea of the user interface, are very welcome. Please just leave a message in the forum.

BibWorm is currently undergoing major structural redesign. The goal will be to make it much more plugin driven, allowing the development of a wide variety of add-ons ranging from such tasks as interaction with desktop search applications (e.g. Google Desktop Search or Lucene) through database access to custom user interfaces. Review also includes the necessary infrastructure to access online databases and library catalogues via Z39.50 and SRW/SRU. Code published here will, until further notice, continue to constitute tech demos. In the process of this redesign, BibWorm might have to be rewritten in a different programming language, although I am optimistic about someone being able to run it on Mac OS and Linux using the Mono Framework. We will see...

BibWorm currently uses ADO.NET's DbProviderFactories for data access. The test database contained with the sources / release is an SQLite database. BibWorm uses the SQLite .NET data provider (available at http://sourceforge.net/projects/sqlite-dotnet2\). Please install it into the Global Assembly Cache and follow the instructions provided in the help file in order to modify machine.conf accordingly.

I was just reminded by some people that the solution cannot be opened in versions of Visual Studio prior to 8.0. So you might want to download Visual Studio 2005 Express from Microsoft at http://msdn.microsoft.com/vstudio/express/. Alternatively, you can open the solution file with SharpExpress (http://www.icsharpcode.net/OpenSource/SD/) or monodevlop (http://www.monodevelop.com/Main_Page). Both are reported to work fine.

The existing code for BibWorm has been uploaded. It includes plugins to communicate with Google Desktop Search (fully working) and Lucene.NET (not working yet, weird error about a null string that isn't null...). Instructions on how to enable them are contained in note.txt in the trunk directory of teh subversion repository.