From: Jim Cole <jim_cole@Merck.Com>
  To  : rasmb <rasmb@bbri.harvard.edu>
  Date: Thu, 10 Apr 1997 13:08:14 -0500

binary files

Dear RASMBers:

Here is another perspective on this lively debate over file formats. At
present, I are happy with the current ASCII file format with respect to
disk space and data loading rates for our current needs. However, I suspect
that in the future we all will be collecting tons more data using the
interference optics and the ASCII format will cause us grief with respect
to disk space and file transfer times.

I think that there is no problem with the XL/A-I storing data in a binary
format as long as there is an easy way to extract data into ASCII files for
analysis or transfer to other software environments or platforms. Many
analytical instruments store data in a proprietary binary format and supply
conversion routines with their software. Alternatively, I have seen lots of
subroutines written for the Macintosh program IGOR Pro that allow direct
loading of various binary formats. If a binary format is chosen, I think
that it should be sufficiently generic to allow someone with a moderate
degree of computer savvy to write a bit of code to directly load it into
their favorite software.

I am considerable less happy with the current directory tree structure.
Even with just absorption data, it is a real nightmare to keep track of
data from a single experiment when it is stored over various subdirectories
and the file names only indicate which cell the data came from. Tom Laue
brings up some good points about going to some sort of database structure
in which the scans are stored in a "packed" format which can be sorted and
indexed using several data collection parameters.

Jim