I've already replied to George in a private mail, but there seems
to be more general interest in this matter.
Yes, we're actively using Unitree as a MR-AFS residency. MR-AFS
was developed at PSC (mostly by Jonathan Goldick and Bill Zumach,
now both at Transarc). Since Jon and Bill left PSC, there was
no active MR-AFS work there since last years summer. Hartmut
Reuter from IPP Garching, Germany has added some real important
bug fixes and extensions to the MR-AFS code base recently.
We are using the system in production since November 94 on
a Sparccenter 1000, now with Unitree 1.9.
We had much more problems with Unitree (and we still have some
of them...) than we have with MR-AFS.
I gave a talk on last years Decorum about our use, you should
find my slides in the proceedings.
Axel
