High there, some thoughts to HDF 5 first: - the situation is a mess! I, and other people, have already got a lot of NeXus files. However, with the advent of HDF5 support for HDF4.* will cease eventually. So we have to support HDF5 in the long run. - I suggest reimplementing the NeXus API in terms of HDF5 when HDF5 is released with attributes. We definitly need to increment the version number then or give the functions a new prescript, perhaps NX5. - Then HDF5 has a new feature: it supports structures in a form called compounds. This could be used to improve standardization: a building block becomes a compound in the NeXus file and all the data for a component can be read in one go. This has the drawback however, that automatic conversion between old NeXus files and new NeXus files becomes extremely difficult if not impossible due to different concepts. And we loose in flexibility when somebody wants more data for a building block the foreseen in a NeXus component compound. So I rather suggest to ignore this new feature. - I think we will have to provide an automatic conversion tool from old HDF to new HDF NeXus files. - At that time we might want to get a summer student to do the coding. Concerning Stefans F90 issues: - Thank you for providing that interface! - The passing by reference issue: I suggest that you provide a set of C-wrappers for the NeXus functions which do the trick. Two reasons: * I already have a lot of code which would need to be changed if we change that API again. * Why clutter sensible ANSI-C concepts of variable passing for F90 compatibility? A pass by reference C interface would not feel natural to C and C++ programmers. - You are right about the NXhandles, leave them alone. They could be replaced by F90 pointers but only if F90 provides some sort of generic pointer. I'am not so sure about that one. We implemented the NAPI as it is in order to be able to tamper with the NeXus file description structure at will without making changes to user code necessary. For instance move it to HDF5. That's my 2c worth of opinion Mark