Lessons Learned

• XINA Markers

XINA Markers

• Managing Marker IDs (aka primary and secondary IDs):
  • Having to manage a master marker list is prone to error
  • Takes a lot of work
  • Quickly creating new markers for ad hoc tests is slow, which can often result in markers not being created at all
• Procs/scripts that end early, either intentionally or due to error:
  • Leaves dangling open markers
  • Open markers had to be manually closed by knowing the primary and secondary ID of the open marker(s)...error prone
  • No immediate feedback on which markers are opened. Test conductors are unaware that markers are left open.
  • Added convention where an ID of 0 with close flag will close all open markers
    • Doesn't play nicely with long duration "wrapper" markers, since those will be accidentally closed
• Tagging "bad" markers
  • On OCI we use the convention of having a "debug" metadata key. "yes" means the data should be ignored, else "no"
  • Does not work when the proc is automated since the proc will automatically close the marker
  • The 10 minute import frequency means you have to wait to be able to tag a marker as "bad"
  • Not well integrated with the web interface
    • Needs easy filters to ignore "bad" markers
    • Display "bad" tag on markers
    • Need button to easily tag a "bad" marker
• Marker label should be increased from 64 bytes to 128 bytes?
• Metadata
  • key search is case sensitive
    • Solution: enforce all uppercase or lowercase keys