Feedback for HDF5 2.0 development

I started this post for an on-line discussion of the development of the HDF5 2.0 release. Please refer to the official planning documentation: HDF5 2.0 Planning · HDFGroup/hdf5 Wiki · GitHub

I previously posted to the HDF5 2.0 wish list topic. It is really good to see the upcoming support for complex numbers. Now, here are my very few burning issues for the upcoming release.

  • Enable independent building of language wrappers, Fortran in particular. Ideally they should be independent packages.

  • Do not drop 32-bit support unless its problems are insurmountable. This was listed in the October 15 slide presentation. HDF5 has become core functionality for many software packages. It is important to retain functionality for possible 32-bit embedded systems, as well as older conventional 32-bit systems.

  • Keep the planning document at HDF5 2.0 Planning · HDFGroup/hdf5 Wiki · GitHub up to date as a master document. For example, the loss of 32-bit support is not listed anywhere, as far as I can tell.

3 Likes

Hi, @dave.allured !

Thank you for your feedback!

FYI, we have a CI test for i386:

hdf5 dev CMake i386 · Workflow runs · HDFGroup/hdf5