[Edserplo_users] Another major update

[Steve Loch]

This update addresses quite a number of reported bugs, some requested enhancements, some major re-adjustments of the internal code as well as adding some much needed functionality. The reported bugs will have been notified as fixed at this update. Most of these changes are down to Rashid's efforts. 

This should clear the way for Dataring operational use although I would prefer to have the stats. checked out first.


Appearance
-------------------
Most obviously the font and presentation of the Series and Port pages have changed.

The axis annotation has been improved  on the map and elsewhere - by introducing a new class.

The sizes of the "frames" are now  individually adjustable. Many people no doubt expand the frame to fully occupy the window but if you want to avoid a flicker problem with the cycle window it is best not to do this with all frames. You can sit the cycle window outside the plotting frame and then you won't get a refresh flicker each time the screen is updated -  without, now, impacting the display aspects of the Series page for example.

Metadata
---------------
Metadata is now present (this compensates to some extent for the lack of Serplo's Page 2) on SeriesTimPage - as a demonstrator - linked to ZESH tables (see below). It needs to be extended to manage CRP data. Perhaps Gwen and Robin can advise me on the fields required. The idea is that we navigate from the file name to the event number using CTD_MAP, for example, but as the latter table is relatively new this may not be satisfactory. The metatadata is displayed on the CTD page and Timeseries page. It might be sensible to include these displays on the Series page as well


Edteva stuff
-------------------
The PortPage display now features null lines as an aide memoire for ports without data when porch files are specified.

Metadata mentioned above support could be extended to PortTimPage (and PortPage) where it might have some details relating to Ports. EAM please advise.

Apart from the axis code already mentioned, examples of internal code adjustment include the Oracle access code for Edteva functions  which has been centralised. It searches in the current directory and then the $HOME directory for the .edteva.oracle file (the latter ID is defined for Windows as well as Linux - viz C:\Documents and Settings\<id>). The column colour in CycleWindow now tracks the settings colour . This required some re-organsiation of the cycle window code.

The order of Analysis constituents is now properly defined on the Analysis page. 

The OutputPage rewrite has more or less concluded. Hourly residuals now match Edteva's. This has been tested  ...
The ability to generate high and low waters and MSL is there as is the ability to support span dissection and sampling. 

The ability to undertake series timeshifts (as in moving to a different timezone) and get the derived channels to be re-computed when the series in question is accessed from a dump file is now available (this required extensions to the dump file content). In fact this capability seems not to have been present in Edteva.

Duxmod is now available as the script ~bodcsoft/scripts/duxmod to sit alongside edserplo. Those who don't know what Duxmod does needn't worry (direct update of series from dump files); those who do may or may not be glad. It should mean the end of Edteva's activities for GLOSSactivities (but see What Remains below). It doesn't require aliasing information because this is now recorded in the dump file.


What remains? 
----------------------
Buds file locking (on new form only)
Validating statistics
Tidal analysis  
Plotting against different tidal analyses
Possible replacement of spline with a better approximation to Edteva's NAG routines. There is a switch in the code to allow direct access to the native Fortran version. 
Further enhancements to metadata as noted above
Coding an update module for QXF (jury oiut on whether this is actually needed)
Rewrite uk.ac.bodc.netcdf.RandomAccesFile so it no longer has deprecated elements.
Revise axis positioning on inset track map

I have left the Tidal Analysis till last as it has a lower priority than getting operational the functionality needed for Dataring processing. I spend most of my time validating changes (both Java and Matlab), suggesting improvements, solving bugs and producing updated versions.

Further work envisaged
-------------------------------------
Repackaging (part of customisation)
Usage logging by accession, series, cast etc.
Pre-emptive flagging.
Use of web services for paremeter description
Further work on maps
Tutorial