Subject: My notes from PWS meeting  2/1/00

	Meeting Notes
Private Athena Workstation Discovery
	1 February 2000

Present were, kretch, ghudson, sly, miki othomas, ted, rbasch, wdc,
orbitee (Christine Moulen), jcwhite (James C. White), tfitz (Tom
Fitzgerald), and jmilgram (Jerome Milgram).

Summary:

We went around the room and everyone introduced themselves.

Then we went to the people who had identified themselves as workstation
owners and asked them what the did and did not like about the present
Athena Private Workstation.

After that we established the format for future meetings:  Miki
apologized for not sending out an agenda.  Future meetings will have
agendas, and people on the mailing
list will decide if they feel they need to involve themselves in the
posted items.

Miki will post notes and invites people to supplement what she sends
with corrections or additions.

Next meeting will be in two weeks, same time same place.
	Tuesday, February 7, 2000
	3:00 PM to 4:00 PM
	E40-316


----

Detail:

First up was Jerry Milgram who said he'd misunderstood the group's
charter, and that he was looking for a group that was porting NT to
Athena.  Before he left the meeting, he gave some excellent insights
into how some perceve Athena:
	Athena has not offered enough bang for the buck on numerical calculations.
	Although the SGI graphics have been good, he perceives NT boxes as
	a much cheaper way to get his graphics, and number crunching jobs done.
	He perceives the Windows user interface to be much easier than UNIX
	for non computer scientists to use to get the job done.
	He asks if there are word processors and spreadsheets equivalent to what
	NT has for UNIX.

Next up was James White who works in IS on SAP stuff.
	The Athena systems he has are the few surplus Suns in his lab that people
	troubled themselves to power up.
	He finds going to MicroCenter, buying a Linux CD and making that run
	a much more tractable task than getting a Solaris release or digging out
	instructions on how to install Athena.
	He wants to feel like he has a clearer sense of what is happening
	on a system and to exercise more control over what gets installed.
	He echoed Jerry Milgram's sentiment that performance is an issue.
	He finds that the kerberized login gets in the way when all he wants to do
	is turn on a machine, and do a little work.
	He asked for the creation of a clearing house of old equipment so he
	could bring in a bunch of old Suns and trade them in for a current PC.
	Other things that he said helped us distill and sanity check the themes below.

Christine Moulen, who works in the Libraries Systems Office was next:
	The Libraries take care of Unix systems some Athena some not.
	The primary Athenized systems are web servers exporting AFS files.
	AFS is very useful on these systems.
	A key decider between Athenizing or not is whether the 3rd party
	software vendor requires a particular configuration or revision of Solaris.
	Christine likes how Athena keeps patches up to date.  But sometimes
	updates need to be postponed in order to stay at a constrained OS revision,
	Christine reports that documentation on how to do things with the Private
	Athena systems has recently gotten much better, and that there are still
	some improvements needed.

Tom Fitzgerald was next:
	His lab has desktop systems that are used by students.  Those are 
	configured as Athena Public.
	His main use of the Athena Private configuration is for servers.
	His lab has BIG storage needs and loads special hardware and drivers.
	He likes how easy it is to clean up or reinstall an Athena system.
	He also likes how OS patches are kept up to date without his doing the
patching.
	But he doesn't like how updates trash his drivers.
	Some discussion ensued about how this could be minimized, but with the 
	recognition that major OS upgrades, even independent of whether or not
	the system was Athenized would require re-installing drivers.
	Tom wanted help in getting more bugs through to the vendor from us.

Finally Angie Kelic told of her tenure as a sysadmin of Athena systems
in Aero Astro:
	Those systems were the last systems to be timesharing hosts with X terminals.
	It turned out to be easier to install stock Solaris than to try and undo
	the Athena update's trashing of the X terminal support.
	The systems DID have the benefit that they supplied a lot of Athena 3rd party
	software to students.
	Athena and Solaris are perhaps too tightly integrated.
	This seems to end users less to be the case for SGI.

From these discussions the following themes emerged:
	Users find it difficult to find out what they need to know about Athena.
	Athena updates stomping on customizations is a real problem.
	Third party software installation on an Athena system is painful.
	Private Workstation owners want to feel like they are more in control 
	than Athena has permitted in the past.

There is a customer outreach issue that is independent of the Athena
Private Workstation Product.  Some users simply don't know how to apply
what is already available.

Bill expressed the opinion that the tradeoffs done for Linux Athena
answered the technical issues rasied by these customers:
	Using a normal OS installer rather than track makes it easier not to stomp
	on customizations.
	Installing stuff locally rather than pointing to it in AFS helps ease
3rd party
	software installations.