Currently, the hdf5 operations for object serialization in yt are handled by pytables. However, pytables is not the only python hdf5 module out there. For my own non-yt code, I have been using h5py ( http://code.google.com/p/h5py/). For simple hdf5 i/o, I find h5py to be far more intuitive and flexible than pytables. If you want to do hdf5 in python, I strongly recommend h5py over pytables. It's for these reasons that we're moving yt from using pytables as a data-serialization backend to using h5py. Apart from casual use, h5py offers additional benefits over pytables that are relevant to yt. h5py is faster and relies on fewer python object than pytables. h5py works better with yt in parallel and its design is generally better suited to yt. Switching dependencies is something we want to do as little as possible, but it seems worth it in this case and will most likely continue to pay off as yt grows. This evening, we will be committing the switch from pytables to h5py dependency to the yt trunk. We have altered the install script accordingly so rerunning that will do everything that is needed. h5py can also be easily installed on it own. Set the following environment variables: HDF5_DIR=path to hdf5 HDF5_API=16 Then do: sudo easy_install h5py (You may not need to use sudo, depending on where easy_install is installed.) None of the yt function calls have changed. Only the internal calls to tables functionality have been replaced with h5py calls. Additionally, it should be noted that this does not affect any of the dataset i/o as this is handled by Matt's specially built hdf5 reader. So far, we have tested this on various machines and it seems to be working. If anyone encounters any problems that may be related to this or has any problems installing h5py, please contact us as soon as possible. Britton Smith