A data format to encompass all formats - myth or reality?

C.M. Moreton-Smith
CCLRC Rutherford Appleton Laboratory

In scientific computing it has become "fashionable" to devise data formats, usually one or two per field of interest, examples being NeXus, CIF and imageCIF. Although this definition work can give some real benefits, the fact that the attempt at unification produces more data formats than there were before seems something of a paradox.

This paper asks, and also provides some possible existing solutions to the question:
Is there a more generic approach to storing scientific data which does not involve creating more mutually incompatible data formats ?

(posted 21-Oct-97 jw)