Re: file naming

[<aah@nofs.navy.mil>]

Well, crud.  It must be Friday or something. :-)
9x3 is certainly not what I was aiming for.
Thanks, Tom, for catching that.
  So we have:
     jjjjfxxx.fit
     
  where
    jjjj = last 4 day digits of the Julian Date
    f    = filter name (bvri)
    xxx  = sequence number starting at 000
    
with files stored in site-named directories.  When copied
on hard disk, files must either be inserted into equivalent
site-named directories, or must be named by concatenating
the site name with the raw file name.
  Since we are currently at JD2452075, I don't think there is any
need to worry about rollover in the last 4 digits; we have
about 8000 days or 20 years before that would happen.
  The one remaining issue is whether the extension should
be .fit or .fts.  My preference is .fts, but someone who
is more familiar with the Micro$oft world might know what
extensions are most commonly used.  I remember a conflict
with one of these.
  The problem with Paul's scheme of using other characters
in the filter field for designating bias/dark/flat is that
B for bias is too similar to 'b'.  While the current Mark IV
systems are only using V & I filters, there are at least
two future systems that will use B & R.  One could, for
example, just increment the filter letter by one for flats,
changing 'bvri' into 'cwsj' and add, say, y for bias and z for
dark, but that might be getting pretty confusing.  Again, what
we are identifying here are the raw frames.  Once they are
processed, you would have to use some unique 'processed'
identifier.  That is why the 8x3 format with the sequence
number as the extension is so nice; it frees up three characters
in the base file name that can be used for such niceties.
Arne