Present: Amber Boehnlein, Ashley Deacon, Richard Dubois, Thomas Eriksson, Stuart Marshall, Randy Melen, Nan Spiers
DLP: Nan
The Symantec DLP PII scanner was disabled last week, pending a re-evaluation. Nan will be leading the effort to try again and will be sure to go through the project process and get guinea pigs from the science community to try it out before re-deployment. Her feeling was that the existing tool can do the job; it just needs to be better configured before letting it out in the wild again.
Hardware for FY12:
PPA seems to be doing most of the buying. Amber has the list. We are still querying AD and LCLS to see if more might be coming.
A question came up in PPA about updating the interactive login severs - the norics and yakuts. The latter are SL6 machines, and suspected to be only used by PPA. Amber will conduct a survey to see who uses the norics (and yakuts) before the next steps of an upgrade path and funding scheme.
Review of CD:
The next notable step for the SCSC is the review of the computing division's scientific computing activities. This has been slated for the fall timeframe. We want this to be a positive experience, educating people in what CD does and trying to see what could profitably be done differently.
There was some discussion about how to do this in a review process - one in which we plan to have external reviewers. One suggestion was to have a pre-review, say in September, where the CD folks give a fairly high level review of their current practices, then use this as lead in to a more in-depth review with the external reviewers later in the year. No conclusion yet. There are already constraints in September, so we'll need to nail down any dates soonish.
Scientific Computing Processes
While looking over the list, it was felt that we should first identify what we expect to happen in scientific computing: who the players are and what their roles and responsibilities are. This turns out to capture the process list that we showed at the workshop.
A draft of the 'execution plan' has been started and will get fleshed over the course of a couple of weeks.
Recap of Workshop Postmortem - Amber
The local organizing committee met to discuss the workshop. Turnout and the caliber of the talks was very high. Much of the discussion centered around taking the next steps to keep the momentum and interest high. The format of the short talks worked well to give an overview, however many topics could be covered in more depth. One idea is to form a working group (or more than one) that would meet regularly on topics of interest to share expertise, and presentations on how to do things--generally or in the context of SLAC. Topics suggested include GIT and other development/developer tools, GPUs, mySQL, frameworks, Data management. There was also further discussion of using and encouraging the use of the forum (see for example, the GIT thread https://forum.slac.stanford.edu/index.php?t=msg&th=20&start=0&). An early topic for a working group discussion is GPL and other licensing issues and tech transfer. A forum topic under 2011 workshop has been started to get your ideas for the working topics (or prioritize the ones listed). We will schedule the first working group meeting for mid/ late July.
We discussed following up on the issue of SLAC having an INDICO server. In this area, it would good to understand the overall SLAC strategy for web content/web management and so a representative from communications to address the SCSC on this. You can join the discussion at https://forum.slac.stanford.edu/index.php?t=msg&th=11&start=0&
Miscellany
Thomas noted that there are no links evident in the SLAC web to confluence or the forum. You need to know where they are to find them. We should figure out some good places to link in to them. [I wonder where sharepoint gets pointed from? (I wonder where sharepoint gets pointed from?)
It was felt that the forum tool needs easier notification schemes and that success will depend on people being told that there is new content, rather than expecting them to browse the forum regularly.