(Summary comments indented after original agenda items)

Present: P. Fritschel, G. Gonzalez, B. O'Reilly, R. Rahkola, K.Riles, P. Shawhan, D. Sigg, S. Whitcomb, J. Zweizig

* Collecting & summarizing segment information - general approach
  - How best to solicit and collect info
  - Format(s) for presenting compiled info

KR: Need to spell out for search groups which segments are truly problematic. Can't expect them to make judgements on particular filters, gain settings, etc.
SW: Should distinguish between effects that increase noise (automatically detected by search groups) and ones that affect calibration or cross-coupling coefficients. (Cross-coupling changes could invalidate veto safety studies based on other data segments.)
KR: Will generate graphs (format not yet decided) to give visual overviews of segments and problems, day by day.
JZ: Can add segment labels to summary plots from DataQual monitor
All: For now will produce tables of segments for posting, as opposed to rewriting SegGener database segments (since those seem to be used by online searches only)

JZ: Rerunning DataQual, SenseMon & glitchMon over pre-raid-crash data on Caltech's saiph machine. Currently processing data at same rate as original data collection. Working on speeding things up. Any needed future reruns looking at RDS channels should be much faster.
GG: Notching out L1 calib line above 900 Hz? (dominates high-band RMS)
JZ: No, but can sum several smaller-range bands to avoid the line - will look into it.

SW: Daniel & Rana should look over filter change list to judge whether segments with non-standard settings should be tossed.
DS: The H1 & H2 michelson filter changes may be benign to GW channel, but hard to say with confidence. Might be safer to veto affected segments.
KR: Will make sure segments are flagged in vsn 2 segment list and quantify data lost.
PS: For technical reasons, the change in L1 PRC was not flagged in vsn 1 segment list.

GG: Hope to complete L1 list this week. Will take a look at H1 & H2 lists, with L1 experience in mind. Will try to recruit volunteers to help, if needed. Will revisit L1 list and try to clarify which problems might meet Stan's criteria of changing calibration or inter-channel couplings.

SW: Betsy Bland at LHO reviewing H1 & H2 elogs (detector & operations)
BO: Joe Langdale has looked at conlog changes  already and will be going back through elog. BO will give guidance on what to watch for.

GG: What about 1/16 second DAQ dropouts (AS_Q = 0 for 1/16 second) seen during S2?
JZ: BitTest should have seen all of those, but the number reported by observers seems larger than BitTest logs indicate. Will look in detail at some examples.  May be necessary to run dedicated monitor over RDS.
JZ: Occasional 1-minute intervals with bad SenseMon inspiral ranges may be due to lost frames when monitor cpu fell behind. Will rerun SenseMon for those intervals, if needed.

GG went over Rana's list of worries and explained what wasn't clear. Most items do not yet have a volunteer looking at them.
KR: will try to draft prioritized version of list and circulate to this group for review before posting for detchar and search group consideration.
JZ: has some old simulation code that can be resurrected to look for hidden saturations in data chain. Had put simulation on back burner last year when troubles with undocumented filter constants uncovered. Documentation thought to be much better now.

BO reported seeing similar trouble in L1 data as seen by RS when LSC processor fell behind in synchronization.
The AS_Q "clipping" RS finds evidence for may be an example of the hidden saturation worrying Rana.

KR will encourage glitch investigation team to tackle the high-amplitude glitches early on. The glitch team has strong overlap with the burst/inspiral veto team which will be encouraged to focus on understanding the vetoes found to be most effective in killing burst candidates while maintaining low deadtime.