Talk:Marks - DatLab

From Bioblast
Jump to: navigation, search

DatLab 8 perspectives

  • The task is to save several analyses of a recorded signal, e.g. have a protocol analyzed several times in the same file, where it would be possible to set marks differently.
Setting Marks is one of the main aspects in DatLab analysis, with different stragic options:
  1. Return to a concept followed in DL2: Recorded files had a specific extension (*.DLR), which changed after off-line analysis to *.DLA. Thus any new analyses never touched the *.DLR file, but different analyses could be stored (together with the raw data). Limitation (1.1): different analyses could not be compared in the same file. This may be acceptable, since the results can be compared in a mark statistics analysis. Limitation (1.2): Anybody could re-analyze and overwrite the *.DLA file, such that the history was lost.
  2. Continuing from DL3 to DL7 and DL8: The recorded data are always retained, but events and marks can be changed without historical track record. This makes sense, since additions, deletions and corrections of marks (and even events in special cases) may be required as a basis for all meaningful further analyses. Limitation (2.1): Traceability may be too much compromised.
  3. DL8 introduces and saves analysis versions (v001, ..) of marks and events, thus always preserving the original raw version of the file (marks and events) saved on-line, and making versions traceable. Limitation (3.1): Like in (1) and (2), this works only, if all analysis versions are saved in a single data storage room, where version numbers are added automatically. Otherwise, the version numbers will diverge on different computers or different servers, without solving the problem.