Versions Compared

Old Version 9

changes.mady.by.user Unknown User (chevtsov)

Saved on

compared with

New Version Current

changes.mady.by.user Debbie Rogind

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Migrated to Confluence 4.0

Currently, you can run the CSS DataBrowser in production by going to LCLSHOME => User Dev Displays... (top right corner) => CSS DataBrowser... (top left corner)

CATER Requests that address maintenance/upgrades to the "old" Archiver:

  • CATER Request 95730, LCLS Archive Engine Maintenance, which refers to work on the current (soon to be retired, "old" archiver) has been closed, and any pending tasks have been moved here (3/2012) in anticipation of the new archiver being online this year. We don't want to put too much effort into fixing the older archiver at this point.
    • Modify aidaGetHistory.m and plotHistory.m to keep track of valid PVs a user tries to view but are not archived.
    • Add Help/guidance for rules on archiving a PV (on change, basic use cases, what to do if a formula is used when some of the data is saved at 1Hz while others on change, etc?)

This page tracks issues with ArchiveViewer.

Bugs

  1. The dock/un-dock plot panel didn't work well. When un-dock, the window doesn’t have title. When dock, the order of the panel tag is changed.
  2. I just tried a simple X1 - X0 formula for two temperatures in the injector. Errors seem to indicate that the two data sets were of different length (very much expected given the way the archiving works). The program should have an option to interpolate data so that one can do A-B.
  3. I haven't been able to get the Formulas to work at all. It shouldn't be offered if it doesn't work reasonably well. I feel I wasted a lot of time trying.

Requests

  1. Inform why a formula failed
  2. "JFreeChart for Time Plots" etc. tab names are meaningless to me and Correlations and Waveforms functions are not implemented. Please don't use controls jargon on these user displays if nobody gets anything from it.
  3. It would be more useful if the visibility box wasn't buried inside the ellipse box.
  4. Non-intuitive controls (how do I select dots instead of connected lines?)
  5. EXPORT limits exported data to 50 hours at 1 per second. This needs to be a lot larger. 1 month of data is a reasonable maximum limit.
  6. Drag or copy/paste mostly works. It would be nice if Archive viewing were more integrated with EDM displays and not a completely separate application. (EDM Action feature).
  7. If time range spans multiple Archiver indices, user needs tell the program where to find the data.
  8. There is no math in correlation plot (correlation coefficient, polynomial fit, psd, etc...)
  9. Integrate with live data (e.g. show current PV value, .EGU and .DESC)
  10. After a PV is placed in the input window a CR launches a search window. Almost always, because the PV is dragged into the window, there is no need for a search and it would be more appropriate to have CR add the PV to the plot list, and use the ADD tab for the search dialog.
  11. Often values for the data being plotted appear with too large a numbers of displayed digits. For example plotting a quad magnet current, the Title is " QUAD:IN20:651:BACT (-251.2770000:20.4479000kG) " . We don't have any numerical values that have 10 significant digits. Generally if we can't see the least significant digit it doesn't matter. Four or five significant digits is plenty to display.
  12. The time axis labels are too verbose. A 24 hour plot does not need to have the day month and year at each label. Also "PST" in the Time axis label is unnecessary and distracting.
  13. Show min, max, std, mean, min-max and std/mean
  14. Option to show limits scale with lines indicating the limits.
  15. Option to median filter the data.

Training

  1. When plotting "long" time spans, it seems to make something very strange: Instead of plotting 10000 points / mm, it wants to be "smart" and plots only three points: the middle (or mean) and the max and min during that time bin. This distorts the data, especially when you are looking at phase variation of a few degrees and once in a while there is an erratic measurement of +180 or -180 deg. This show up as a "nicely" colored block with many vertical lines in there making the real data nearly impossible to see.
    1. Even more annoying is some temperature data. For reasons I don't understand and are not due to the archiver program, some temperature values are reported to be very large numbers that have at least 32 decimals to the left of the decimal point. This really screws up the plot scale. Such numbers should be ignored when setting the plot scale. Better yet these numbers should be purged.
  • Sergei's comment: Right-click on the value => Select "ignore value"
  1. Very difficult to arrange different scales for two PVs on the same plot (required months of training)
  • Sergei's comment: Actually, this takes only 10 secs if you look in the menu bar under Tools (wink)