karen@informatics.rutherford.ac.uk (02/25/91)
In ISODE 6.0 FTAM, the FADUidentity structure allows a FADU to
be specified by name or list-of-names. In each case, simple
null-terminated strings are assumed.
As far as I know (and I'm sure someone will correct me if I'm
wrong on this!), there is nothing in the FTAM standard which
restricts node-names in the VFS to such easily handled objects.
One could imagine a document type where the logical choice for
a name was a particular chunk of (binary) data from the file,
embedded nulls and all.
This in fact almost happened recently for the emerging FTAM-6
document type (for handling of Computer Graphics Metafiles).
When it was decided to stick to printable strings, I heaved a
big sigh of relief, as I am involved in adding support for FTAM-6
to ISODE as part of the ESPRIT II project "ARGOSI".
Now I am wondering, although I don't need it immediately, whether
it would not be for the long-term best if ISODE's FADUidentity
structure specified names as bounded strings (a "char*" and
an "int" giving the string length) rather than just "char *".
Up till now ISODE has not made much use of FADU names, so if
such a change ought to be made this would be an excellent time
to do it. What do you people think?
Karen
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Karen Goswell DCSD, Informatics Department
Usenet:{... | mcvax}!ukc!rlinf!karen Rutherford Appleton Laboratory
JANET: karen@uk.ac.rl.inf Chilton, Didcot, Oxon
ARPA: karen%inf.rl.ac.uk@nfsnet-relay.ac.uk OX11 0QX, UK