Changeset 656

Show
Ignore:
Timestamp:
09/30/08 14:55:56 (2 months ago)
Author:
laidler
Message:

Updated notes.

Files:

Legend:

Unmodified
Added
Removed
Modified
Copied
Moved

  • trunk/pysynphot/planning/pinned_throughput.txt

    r655 r656  
    4343throughout the system. 
    4444 
     45WHICH INSTRUMENTS ARE AFFECTED? 
     46=============================== 
     47Only a preliminary estimate has been made by checking for the 
     48existence of throughput files that are not precisely equal to zero at 
     49both endpoints. This is an overestimate, because some files, such as 
     50longpass filters and internal optics, should not in fact fall to zero. 
     51 
     52According to this preliminary estimate, 
     53  ACS, NICMOS, and STIS have many files that are affected. 
     54  COS has a few. WFPC3 has only 2 aperture files. 
     55  WFPC2 has many; so do FOC,FOS,HRS,HSP. 
     56 
     57All supported non-HST modes (Johnson, etc) are unaffected. 
     58 
    4559 
    4660WHY SHOULD WE CHANGE THE FILES? 
    …… 
    6579 
    6680Option 1: Update the throughput tables so they are well defined. 
     81 **This option is preferred by the NICMOS team. 
     82 
    6783  Pros: 
    6884   - gives you complete control over the throughput behavior 
    …… 
    7692      - editing: updating the tables to include them 
    7793      - testing: testing new throughput tables (how?) 
    78        
     94 
     95 
     96How can we minimize the work required? 
     97-------------------------------------- 
     98The NICMOS team recommends the following actions to minimize the work 
     99required by the teams: 
     100 
     101  - In changing the files: 
     102 
     103    - define a standard modification, eg, "add an extra point 0.001A 
     104      from the current end of the table with throughput=0". Then SSB 
     105      could write a script that would apply this change, and the teams 
     106      would only need to provide a list of affected files.  
     107 
     108      It would be possible to selectively hand-edit files that warrant 
     109      more sophisticated extrapolation.  
     110 
     111  - In reducing the testing/validation load: 
     112 
     113    - The algorithm described above, by essentially making a step 
     114      function just past the current end of the defined data, would 
     115      come closest to the current SYNPHOT behavior, which would in 
     116      turn minimize the expected change, thus making it easier to 
     117      validate the test results. 
     118 
     119    - Using the ETC regression tests (ie, the last public release of 
     120      the ETC software, using synphot) as the pre-delivery tests would 
     121      preclude the need to invent new tests. 
     122 
     123    - Establishing an appropriately small difference threshold, below 
     124      which the tests would be validated, would further make the 
     125      validation stage easier. 
     126 
     127 
     128 
    79129Option 2: Add a header keyword that will tell Pysynphot to taper the 
    80130   throughput curve to zero when the file is read in. 
     131 **This option has been rejected by the NICMOS team. 
    81132 
    82133   Pros: