Re: file naming

[Chris Albertson <chrisalbertson90278@yahoo.com>]

If you leave off the site ID then it is very likely files in FTP
/incoming directories will have the same names.

Here are a few ideas to save one character:

1) Why record the filter type?  It's in the header.

2) Use Julian date/10.  So it changes every 10 days then
the sequence number is made base 36 not base 10.
So you count 001, 002, ... 00Z, 010, ... ZZZ.
This allows 36^3 images.

3) Lose the date.  All OSes already date stand files.
Just site ID and sequence number.

4) Here is the one I like:
  Assign Site IDs A, B, C and so on then let every site
  use what ever name system they want so long as it is
   1) An 8.3 lenght
   2) Each file is unique.

The problem with selecting a .fits, .fit or .fts suffix is
that every FITS reader has its own convention.  You can't
pick one that makes them all work.


Sequence numbers

Tom Droege wrote:
> 
> OK, we are close.
> 
> I note that Arne has listed a 9+3 proposal below.
> 
> >ajjjjfxxx.fit
> 
> I would agree to leave off the site from the string to get to 8+3 and use a
> directory instead for each system.  This leaves
> 
> jjjjfxxx.fit
> 
> jjjj  Julian Day
> f      filter
> xxx  sequence number
> 
> I like this better than cutting down the JD by one which was my last
> proposal.
> 
> This would then be the standard file name for data on CD.  There would be a
> directory name, which I propose to be the name stamped on the telescope if
> there is no better proposal.  If more information is needed, then one has
> to read the fits header.
> 
> I am happy to stick with our old site names.  I am H for TOM.
> 
> Arne, do you want to update your old TN or should I make a new one?
> 
> Tom
> 
> At 01:32 PM 6/14/01 -0700, you wrote:
> >Chris is concerned about files being moved around and
> >their uniqueness lost.  The concept was that files on
> >CD be unique with 8x3 character; files on disk can be
> >named anything you want.  Someone (sorry, my memory is
> >poor) suggested concatenating directory name with file
> >name for hard disk storage, which is one solution.
> >   Another option, which gets back to the Mark III days,
> >is to use that first character (which I've set to "J")
> >as a unique indicator of site.  That only gives you 26
> >sites without special characters or discrimination
> >between lower and upper case, but perhaps that would be
> >enough considering the time it takes Tom to finish
> >systems; he might not reach #26!  That is a reasonable
> >compromise to me, with the first character starting at
> >"A" for TOM, "B" for ARNE, etc. in order of completion.
> >   Therefore, the proposed scheme is
> >      ajjjjfxxx.fit
> >      a = unique site identifier
> >      jjjj = last 4 days of the Julian Date
> >      f = filter
> >      xxx = running sequence number starting at 000
> >      fit = file extension for fits files
> >This seems to be the simplest scheme, and might be almost
> >human-readable.  I'd still prefer files to be stored
> >on CD underneath a site directory.
> >Arne

-- 

--
   Chris Albertson             
   chrisalbertson90278@yahoo.com
   Redondo Beach, California
   home: 310-376-1029
   cell: 310-990-7550