This is a big problem for us at Oxford and we are pretty cautious
in our approach to rolling out new client software.

There is a clear case to use the latest client code when there for
new platform support.  It is much less clear that persuading people
who are happy and working well on the old version that life will be
better if they use new code.

It has been our experience that new is not always better ...

An example (I am almost, but not quite ashamed to admit) is that on
my Solaris workstation I am using 3.1.0.7.  A great vintage and it
is working well will TSM 4.1 server.  Ian (my nice colleague) has a
brand new Solaris workstation and is running a 4.1 flavour - sadly,
his scheduler has died several times in the last two weeks with no
clues.  A recent PMR has highlighted that common code in one of the
4.1 versions (we saw the problem first at AIX) has memory leaks.

It is difficult.  What we really need to know is why the 3.1 client
(or even earlier) is not actually going to work with later server
versions?  And is this really true for all older platforms?  As well,
not all our users can afford new kit, new versions of OS and gain a
natural opportunity to upgrade to the latest TSM.

Another issue I am not very comfortable with is releasing client code
which is more in advance than the server code - again, the technical
issues relating to this would be appreciated.  What policy do others
have on this?

Regards, Sheelagh
--
Sheelagh Treweek
Sheelagh Treweek
Oxford University Computing Services
Email: sheelagh.treweek AT oucs.ox.ac DOT uk
Phone: +44 (0)1865 273205 Fax:-273275
+---------------------------------------------------------------------+
|  http://tsm-symposium.oucs.ox.ac.uk/   OXFORD 20/21 September 2001  |
|  REQUIREMENTS http://tsm-symposium.oucs.ox.ac.uk/requirements.html  |
+---------------------------------------------------------------------+