We just had a discussion w/ @kittisopikulm about Java APIs and bindings for HDF5. Besides idiomaticity, ease of deployment in the target environment is perhaps the most important usability consideration. For the HDFql guest in a Java host, this suggests that an HDFql Maven package might deliver a usability boost for Java developers and users. @contact, have you considered this deployment option? Any showstoppers? I believe that expertise is abundant in the community to help such an effort if other commitments or priorities prevent you personally from pursuing this.
The same goes for the JNI bindings; we’ll start a separate discussion on that.
I enjoyed our conversation this morning. Overall, it would be great to see mavenization of HDF5 libraries for Java. I will also point Curtis Ruden here to coordinate efforts.
Yes, it is in our roadmap to start packaging HDFql not only for Java but also for C#, Python and R to ease its distribution and installation. The idea is to leverage from the following technologies to achieve these goals:
We are planning to release HDFql version 2.5.0 (which will use the upcoming HDF5 library version 1.8.23, and include new features and many bug fixes) during the first half of this year. In the second half, we plan to implement the aforementioned packaging.
