From: Jo Butler <pjgb@mrc-lmb.cam.ac.uk>
  To  : RASMB@BBRI.HARVARD.EDU
  Date: Wed, 9 Apr 97 11:58:39 +0000

Data storage format

Dear Rasmbers (and particularly the software gurus),

I have read the suggestions of a move to a binary data storage for the
XL-A/I
from John and Tom with interest, and also the important points made by
Jim.  As
someone intermediate between the extremes of computer competence, I would
like
to add my comment to this discussion.

One thing which we here in the UK stressed to Beckman, when the original
software was being developed, was our desire to have our data readily
accessible
to us, for whatever purpose we subsequently wanted to use it.  Beckman
responded
nobly by giving us the ASCII format and making it possible to recover the
data
file either directly or from within Origen.  While I appreciate the
convenience,
for software development, of using binary files, unless the capability to
convert very easily to ASCII is retained this will lock out a good many of
us
who just want access to our data, without having to write a routine to
rescue it
from a more restricted format.  I would stress that this requirement is
not
merely the backwards compatibility which has been mentioned, but rather a
continuing plea from those of us who do not wish to have to go too deeply
into
the data transformations, but still want some access.

Two of Tom's proposals concern me in particular (using his lettering):
I  File format proposals
 B.  For general access, the proposed format of the radial information is
clearly much less convenient than that currently outputted by the
instrument. 
While I do not know the details, I rather expect that it might involve a
conversion from the form currently available - which would then have to
be
reconverted back if one wants 'true' radial values.  This is inefficient
and
prone to introduce (at least) rounding errors, but to what advantage when
any
programmer wanting the alternative format can carry out the necessary
single
conversion.
 E. For a large majority of users, optical density is much more useful
than
intensity and we currently have hard/firmware provided by Beckman with
the
instrument to give us what we need.  There are, of course, occasions when
one
wants the more basic intensity data, but this can be obtained easily as
things
stand.  Anyone who has used 'radial.org' to convert from intensities to
ODs will
be aware how slow this is - and therefore how much time would be lost in
abandoning the current conversion facility within the XL-A board.

Please do keep up the software developments, but be careful not to throw
out the
'baby' of convenience for more basic purposes to get the 'bath water' of
a
system which is only helpful for the most advanced analyses.  Anyone
conducting
such analysis is likely to have the fastest computers, together with the
largest
disks, and it would be much easier for them to carry out any conversions
and
store extra files than it would be for us simpler folk to have to work
backwards!

Jo Butler
Jo Butler, 
MRC LMB,
Cambridge, UK.
Tel:+44 1223 248011
FAX:+44 1223 213556
Email: pjgb@mrc-lmb.cam.ac.uk