EPICS V4 Telecon 11-Jun-13

======================

See minutes below

Agenda as Followed

----------------------------

0. Preliminaries (5 mins)

1. Data n-type cont'd (Bob, 30 mins) -> 2nd TOPIC

 First proposal for general dimensional data type. See Bob's email and

 enclosure [2]

2. Convert discussion, if needed -> was not covered

 Again there has been a lot of discussion on the list. Last week we

 had agreement (although no RESOLUTION) on the proposed design, so it's

 going ahead that way, but let's have some tome to discuss issues verbally.

3. 4.3 beta planning [2] (20 mins) - left over from last week.  -> 1st TOPIC!

 Confirm what will be in 4.3:

    easyPVA?

    exampleCPP and exampleJava?

    1 tar or many as in the past?

    Should anything be removed from pvIOCCPP to make it so it doesn't contain dead wood?

 Web site and doc changes for 4.3 beta.

 Packaging 4.3 beta. Who and when.

[1] http://epics-pvdata.sourceforge.net/home.html#usefulinfo

[2] http://tinyurl.com/n38h7kz

   Powerpoint, https://sourceforge.net/mailarchive/attachment.php?list_name=epics-pvdata-devel&message_id=78442E0949CA8C4FA7A701DFFDBC67C315766DFC%40EX10-MB2.BNL.GOV&counter=1

Minutes

======

PRESENT: AJ, DH, GW, GS, MD, SV, TK, BD

SCRIBE:

CHAIR: GW

******

NEW TOPIC: 4.3 beta planning

******

Item 3: AJ agrees to coordinate release.

******

NEW TOPIC: Data n-type (cont'd from last week)

******

BD: Two mirrors. Variable derived from energy. Energy inner loop. Two space variables. Moving sample. Rastering sample in x and y at each position do energy scan. But the order may be different.

Spatial axes are microns. Energy in eV. Didn’t capture this in V3.Can capture this in V4.

GW: Step size? Interval name. Nonmonotonic= “AxisArray” - but INterval still has value?

GW: AxisArray really is all indep variable values.

GW: Only position and time are used as indep variables!

GW: Interval is overloaded.

GW: “frame” = sample id.

GW: Axis description “x”, “y”, “angle”. Needs to be very general.

BD: Tile detectors to make larger detector. Gaps between detectors.

BD: Once get beyond 2 dimensions other axes are non-spatial (time, energy).

BD: Binning integral over area. Axes would change but data type wouldn’t.

GW: Where have you expressed resolution of the detector?

DB: Resolution is spatial so axis info would be different.

GW: No special spatial resolution necessary?

DB: Yes.

AJ: Not specify how big each pixel is, just position?

BD: Not necessary to specify size because know locations and have info on gaps.

BD: Resolution is interval. Does not apply at a gap.

BD: Users want data geographically or chronologically.

BD: Requests:

Give me detector readings at a time.

Give me detector data at all spots at an energy.

Give me readings at one spot at  all energies.

DB: I don’t know of any 3d detectors. Always 1-d over time or 2d over time.

GW; Wouldn’t make them go over time to do correlation.

GW: How do they know how the can correlate 2 frames?

GW: Would have recored energy in data.

AJ: Time is irrelevant to results.

GW: If time isn’t independent varaible you don’t need to record it separately..

BD: How do I find data corresponding to energy.

GW: Need to timestamp data. Correlate over every independent variable that you’re interested int.

BD: Have to store data in every view possible or store raw data and enough data to recreate every view.

What we need to do is the second.

AJ: Are you missing axis name in your description?

BD: Straw man. There are pieces missing. Depth? Angles?

BD: Only every need 2/3 -dimensions.

AJ: Not sure there’s no such thing as N-d array.

DH: Will pass slides onto beamline guys.

BD: Looked at Ulriks data sets and everything is x,y and time.

Any other dimensions they think they have are really not dimensions.

GW: Sometimes time isn’t a variable.

BD: Give me an N-d data set and I’ll represent it as described here.

GW: If this mechanism is meant to represent non-detector data I don’t see data.

BD: Error is something we need to discuss. Applies to detector as well.

BD: Send me you data sets !!!

Post meeting discusion.

AJ; Can I turn on RTEMS.

GW: Good idea. Do initial build first.