[Edserplo_users] Major update - 1.5

[Steve Loch]

The version number of Edserplo as been adjusted to 1.5 to indicate we have made some progress! Once the outstanding items have been dealt with (by mid-year) we'll mark Edserplo as 2.0. Adjustments have also been made to Matlab startup files to accommodate the new NetCDF API. Some enhancemenst made before Christmas by QTL have finally been incorporated.

Changes behind the scenes.
  Use of SVN to manage production code - corrections should be published sooner
  Change to latest version of NetCDF API  
  Change to JDeveloper to 11.
  Change of Java to 1.5

Although the impact of these changes should not be obvious to casual users of Edserplo, the changes are very important for the long-term support of the project (good for 2 years I would guess if there was an embargo on further such changes). Getting to this point has not been entirely straightforward however:  Subversion's integration with JDeveloper on Linux has been a  problem for the last two months as has  JDeveloper 10's  crippling bugs in the debugger!

Flook support (not strictly part of Edserplo but held in same JAR file and uses the same format processing capbility)
  Now implemented through standard JAR file
  Note addition of capability to override listing format (this has to be retorfitted to QXFList)
  Note also that a1d2qxf and ott2qxf are likely to be retired (use flook with option to output to QXF instead).

Output page (Edteva stuff)
 A major rewrite has been going on since Christmas. Some cleaning up remains to be done (i.e. the class will lose a few hundred line sof redundant code).
 Transmutation of flags - implemented through lookup tables so it is fast - still some work to be done in the context of defined-interval output
  High/Low  implemented
  Mean Sea Level  - done
  Statistics - done
  Spline capability to match NAG if NAG library available (not available on Windows)*
  Imperfect support of defined-interval output means that hourly MDL output not yet available 

Dump File (more Edteva stuff)
  Support of recalculation of residuals following a time shift - some of the work done but not all yet

Maps
 Starts at correct scale
  (but axes still leave something to be desired).
  
TimeSeries
   Cycle window keeps its surround
   Line thickness can be changed

Tides
  Elements of have been redone
  Analysis page presents the constituents in the correct order

*The Java spline posted by Samrat is OK for most purposes but it's not so good in pathological situations - such as I encoutered earlier in the year with some Cromer data. The intention behind this work is to provide a Java equivalent for the NAG routines (if deemed worth it). 

Canvassing for enhanecments will occur prior to the IT discussion meeting towards the end of June. However you can always submit a request for enhancement to Bugzilla at any time. 

Steve