I could only give an incomplete answer to Alexej Jerschow's question
about archiving data, but it reminded me of a related topic, namely an
efficient scheme for data transfer, e.g., over the Internet. At the
start, I should admit that this scheme is inconvenient, but it is
worthwhile in cases where transmission time is much longer (or more
costly) than processing time.
Background: The Unix compress command seems to work very well on raw
data, but not on processed data. This has to do with the way Bruker
software uses the 32-bit data words; processed data uses all 32, but
raw data often uses much fewer, depending on the experiment.
The idea, therefore, is to delete the processed data, then archive to a
file, and then compress. For example, in a Unix session, I might do the
following:
cd /u/data/tony/nmr
find sucrose.13C \( -name 1r -o -name 1i -o -name 2rr -o -name 2ii -o
-name 2ir -o -name 2ri -o -name 'dsp*' \) -exec rm {} \;
tar cvf /usr/people/tony/tarfile sucrose.13C
compress /usr/people/tony/tarfile
(That find command is a little cumbersome, so I point out that the
command "delp sucrose.13C" in Uxnmr or Xwinnmr does essentially the
same thing.) This results in a file called tarfile.Z, which might be
transmitted by ftp and processed by the recipient as follows:
uncompress /usr/people/fred/tarfile.Z
cd /u/data/fred/nmr
tar xvf /usr/people/fred/tarfile
The recipient must then re-process the data, with efp or xfb for
example. All other information (for further processing, integrals,
plotting, and so on) is retained from the original processing.
Tony Bielecki
-- ------------------------------------------------------- Anthony Bielecki, Ph.D. ajb@bruker.com Applications Scientist Bruker Instruments, Inc. 44 Manning Road Tel. 508-667-9580, x145 Billerica, MA 01821 Fax 508-667-0985 -------------------------------------------------------