Meeting Notes for 10 December 2008 Upgrade Recovery Helpdesk Form Documentation on Builds for 13 New QuarkNet Site QuarkNet e-Lab Forum for Fellows Hsiao's Report on QuarkNet e-Lab Taiwan Additional Reports that were sent around in e-mail Eric's Report: QuarkNet wiki mashup demo Tom L's Report Cosmic ND Shower Bug Detector calibration efforts LIGO CMS

Meeting Notes for 10 December 2008

Present: Mihael, Eric, Liz, Phong, Bob, Hsiao
Absent: Mike W., Randy, Kris, Joao, Tom J. , Tom L., Dan K.

Upgrade Recovery

Argonne systems upgraded some of the machines and wiped out /usr/local on www13 which was where Eric had some of his software. Someone asked if we should have software in /usr/local (ie build from source) or just use packages? Eric says there are reasons to build both Apache web server and PHP from source, because the packages are not configured the way we need them. He also does not trust Debian versions of OpenSSL since a flaw in how the packages were built was found this year, so he preferes to build OpenSSL from source too.

Eric has created scripts to do the configure and make steps and put these in his CVS to document the process. If there is not a script then it should be as simple as `./configure; make; make install`.

Helpdesk Form

The helpdesk form is not functioning because it was on www13. We discussed where it should be. It requires Apache and PHP be built on the machine. We need a very specific build for the form to work. It could easily go on www10 where he already has everything it needs. We did not think it was appropriate to go on www10, at least now because this is where the new QuarkNet site is being built.

Action Item: Eric is going to to see if he could put it on www18 and bypass www12. Eric has already done builds on www12 and 15.

Documentation on Builds for 13

Eric has extensive documentation the wiki on building the environment he needs. See WebBuild.

New QuarkNet Site

Eric has been working on a new QuarkNet site. Right now it can be accessed by typing quarknet.us. We discussed whether it should be on www10 permanently and if it was more appropriate to have it at one of the QuarkNet locations such as Notre Dame. We can't have it at Fermilab because of the issues of accounts for the users and content on the wiki. Should it be separate from the ELabs servers? If we do dedicate a machine to it on the i2u2 cluster, do we want to ask QuarkNet to buy another machine for the cluster? Do they even have money for that? Mihael did not think it was necessary. We will need to do good backups of this machine because it is very old.

QuarkNet e-Lab Forum for Fellows

Eric copied the e-Lab fellows forums (Cosmic Ray's Diner) from www13 to www10 to get things started. Then TomL posted to www13, (and Bob uploaded to the wiki, which I also copied), so they are out of sync. We discussed where this should belong. It is in some ways related directly to i2u2 e-Lab, yet the members are only QuarkNet fellows and it is private so other teachers using the Cosmic e-Lab would not be able to access it. Should it be on the QuarkNet or i2u2 site? If the wiki goes over to QuarkNet, then the one on www13 should go away.

Action Item: Bob will talk to the e-Lab fellows.

Action Item: Liz will ask Marge where she imagines the QuarkNet site should be and if QuarkNet has funds to buy a machine for it.

Hsiao's Report on QuarkNet e-Lab Taiwan

Hsiao discussed the need for a way to share data between Taiwan and the U.S. We had a long discussion of how this might be accomplished both using the structure of tiers employed by the LHC and a less complicated process of synching data and metadata in the VDC. We also talked about how to implement localization.

Liz: Solving the data sharing and localization issues can mean an large effort and we have to be sure our main goal is to get the e-Labs ready for testing.

Hsiao reported that they use a script in Taiwan to upload data automatically every few hours. We would like to see this. It allows them to upload it to two sites at once.

Additional Reports that were sent around in e-mail

Although we did not get to discuss these topic, I have included content from e-mails received from Eric and Tom L. The latter is important because Tom L. could not partipicate this week.

Eric's Report: QuarkNet wiki mashup demo

I've got the mechanism working on the forums where the content of a page can be edited in the wiki, but shows up on the site as a "plain" page. This is how the Masterclass index page works, see http://www.quarknet.us/masterclass/ As it says there, the body of the page comes from a wiki page that Ken (or Bob or anyone with the right account) can edit, and then it shows up here.

The point being we can use this for ELabs as well. It works.

Another thing to note is that the page you edit is in the project namespace on the wiki, not the main namespace. That means it won't show up in a regular wiki search. That way we keep those kind of pages separate from the main wiki content.

Tom L's Report

Cosmic ND Shower Bug

Having recently confirmed with Tom J that the firmware of the DAQ boards involved in last summer's "ND Shower bug" have not been updated since that time, I'm awaiting notification that any newer firmware exists. If not (and we think not), then the millisecond time lag between older and newer boards remains confirmed but unexplained. If Tom or Bob can confirm that new firmware exists than what we've got installed--Tom J has the details--then we'll upgrade and run the test again. For the record, we carefully ran simultaneous pulses through two old boards and two newer models last summer, and the old boards agreed with one another, but disagreed by a millisecond on pulse arrival time with the two new boards, which also agreed with one another.

Detector calibration efforts

Detector calibration efforts: two HS students are now implementing the calibration procedure which Jeff Chorny and Dale Wiand wrote last summer. The procedure plots efficiency of each paddle as a function of voltage, and then recommends running at the voltage which places efficiency in the middle of the plateau (around 1.0 volt, typically). This procedure can be implemented with only the equipment available to everyone with a QN/Walta CRD. (If you want a little detail: efficiency is determied at each voltage by placing the paddle between two others running at 1.0 voltage, recording the number of counts (on the DAQ's LED readout) at each voltage in triple coincidence for a given duration, and then dividing that number by the number of counts across a similar duration of the top and bottlom paddles in double coincidence.) Once this procedure has been run enough times to determine reproducibility of the curve, and optimal durations (or repeated applications at a given duration) are determined, we'll post the results. We could then invite QN Fellows to comment, and/or directly invite other schools to consider calibrating their detectors in similar ways to enable collaboration on common studies. (We're looking at solar activity studies.) We'll also look at performance studies for detectors calibrated in this way, to explore whether peformance studies might depend reliably on proper detector calibration as described above, and thus could provide a way of determining a detector's proper calibration without going through the efficiency study described above. (The efficiency study produces the efficiency vs. voltage curve, and the calibration procedure says to set each paddle at voltages in the middle of the plateau for each paddle's efficiency vs voltage curve.)

LIGO

Dale and I have worked through and submitted to Jean a rationale for modifying the LIGO content LOs in two respects, as well one of the "getting started" milestones, both of which seemed necessary to pave the way for an adequate LIGO pre- and post-test. Once they've passed Jean's inspection, we can move quickly--we hope still this December--to fill out a draft LIGO test beyond the half-dozen or so questions already formulated.

Another student is preparing a set of questions for Dale, Eric and perhaps Fred Raab or other LIGO scientists based upon the initial results of a 50-quake investigation of the responsiveness of the LIGO interferometer to seismic activity worldwide. He has not yet found a correlation between USGS-recorded earthquakes (plausibly adjusted for magnitude and distance) and interferometer sensitivity.

CMS

Dan's drag-and-drop approach for handling triggers in Monte Carlo data has been implemented in the Test Beam version; it was shown without scaffolding to Tech group last week. Next Monday our team (Dan, Doug and I) will meet to schedule a workflow targeting the production of a posted result of a student inquiry by the middle of January, if possible. (By "posted result" I mean just equipping OGRE with sufficient data and associated scafflolding so that a student can ask a modestly informed question of the data, produce a plot, and write up an account of the inquiry. We could post it to a wiki, or as an e-Lab poster.) I expect to be able to get some of Ike Swietz's time over the break, so am aiming to set him up with something to chew on prior to Christmas, if possible. That entails acquiring sufficient MC data, and making it accessible through OGRE, to support a meaningful inquiry of some sort. Ike has also worked with the unscaffolded new version of OGRE, and I'll get his written reaction.

We are aiming to produce pre- and post-test questions for CMS by the end of January, if possible, for which I need to have some experience of a successful student-level inquiry.

We are also discussing a timetable to introduce OGRE in Jill Ziegler's classroom in PA next semester--early on, we hope--so that we'll have some shot at providing a masterclass pilot activity ready some time in February for a March implementation.

-- Main.LizQuigg - 16 Dec 2008