Matthew Geier wrote:
> I've basicly been told by my wife who has had prior experience with the
>lack to fixed IDs in the past, not to bother attempting to roll out a
>file server that doesnt implement this, and that I should by Zynet(?)
>instead that does do it properly.
When using CAP they had continual complaints about aliases breaking.
>When they switched, the complaints stopped.
I'm not familiar with CAP, but I've used Xinet in the past (a
year ago on an SGI O200). I didn't have a really nice way of
testing FIDs, but this is what I've observed.
There is a program suite that allows the user to generate models
of defined (in a data file) molecular structures. This software
has a companion piece that simulates a spectral analysis of the
molecule. The analysis software can be called by hand, but it
was designed to be called up from within the modeling software.
If the data file containing the molecule sits on a ``real'' HFS
mac hard drive, the user can call up the modeler to display the
structure, then call up the analysis file per above.
If the file is sittine on an AFS volume using Xinet's Appleshare
software, the analysis software starts up, but the data file is
not found. However, the user can still call up the analysis
software and open the data file by hand.
My interpretation: normally, the modeling software opens a file
and obtains the FID, which is then passed on to the analysis
software. However, the FID isn't available, so analysis software
can't automagically find the data file.
Xinet does have appleshare over ddp and ip, as well as pop-up
messages. It also has a set of utils to move and copy files on
the command line without breaking the resource fork.
This was about a year ago, so....
WL
