Keith,

The only significant change made to SpectrumArchiver leading up to M3 and
M4 was expanding the channel list to all PEM channels and all 16 kHz
non-excitation channels. Some typos in the channel name list in the LHO
version of the code were found and corrected during M3. Full evaluation of
SpectrumArchiver wasn't possible during M3 because of a problem with a
stale NFS handle on the control room machines, but Keith evaluated the LLO
version during M4 and made some suggestions for changes to
SpectrumArchiver and the DTT to increase ease of use of the archive.

Two spectra are archived for fast channels, one from 0-1 kHz with a 0.1 Hz
resolution, the other from 0-8 kHz with a 1 Hz resolution. The DTT cannot
view simultaneously the 0-1 kHz and 0-8 kHz
ranges for the same channel because whichever spectrum is read in last
causes the previously loaded spectrum to be overwritten unless the first
spectrum is set as a reference. The fix will be to add the spectrum
resolution to the channel name field in the xml file.

The monitor still runs outside the ProcessManager because as currently
coded it starts once an hour and then shuts down after computing that
hour's spectra. John will be writing a new data accessor method that will
allow SpectrumArchiver to stay up without accessing the data channels
during the waiting period before the next archive time. The new method
will be added to the code for E10.

Finally, the DTT needs to be modified to speed up the navigation of
archive directories. I've been working on this but have been unsuccessful
in testing potential DTT changes because I haven't been able to get the
DTT source code to compile under linux (I don't have a solaris machine at
SKC). Has anyone worked through all the source code changes needed to get
the dtt to compile in linux?