======================================================================

                    H I S T O R Y 

                    doc: Tue Aug 16 11:41:46 2005

                    (c) 2005 A.M. Thurnherr

======================================================================

NB: CHANGE VERSION IN [default.m] BEFORE UPLOADING

IX_0:

    December, 2004:

        - mailed out 1st version with HowTo to Lisa

IX_1:

    Mar 18, 2005:

        - corrected minor bug in [plotinv.m]

    Aug 11, 2005:

        - mailed out to list of LADCP-interested parties as v90.tgz

IX_2:

    Aug 16, 2005:

        - started HISTORY

        - [m/default.m]: updated version, removed date

    Aug 25, 2005:

        - uploaded to website for Lou St.Laurent

IX_3: (REPORTED VERSION IS IX_2!!!!)

    May 18, 2006:

        - implemented bugfix for PPI-filtering multi-ping ensembles

          provided by Mattew Alford [edit_data.m]

        - disabled edit_spike_filter by default

    Jul 19, 2006:

        - [m/besttlag.m] [mlocal/loadctd.m] added code to

          suppress new Matlab 7.2 warning of NaN values in interp1()

        - uploaded to website

IX_4:

    Jan  5, 2007:

        - updated version number

        - merged [mlocal/loadctd.m] from 2 older versions (one with bugs &

          one without support for elapsed time)

        - added support elapsed time-base to [mlocal/loadnav.m]

        - added Dan Torres' code for interpolating irregular GPS time

          series to [mlocal/loadnav.m]

        - added CTD & NAV file layouts to p structure

        - added [mlocal/mkSADCP.m] for pre-processing UH/CODAS-processed

          SADCP data 

        - fixed m/plotraw.m as suggested by Bruce Huber

        - uploaded to ftp site for CLIVAR I8S & Stan's AA cruise

IX_5:

    Jan 17, 2007:

        - updated version to 5beta

        - wrote [m/geomag.m] to use external program to calculate magnetic

          deviation

        - added support for [geomag.m] to [mlocal/loadnav.m]

        - emailed version to Dan Torres & Bruce Huber for beta testing

    Jan 26, 2007:

        - fixed BUG reported by Dan Torres regarding file-layout parameters

          (defaults were in structure p instead of f) [loadctd.m] [loadnav.m]

    Nov 18, 2007:

        - added 3-beam support to [loadrdi.m] 

    May 23, 2008:

        - improved [loadctd.m] to allow removing bad valued

          (f.ctd_badvals) & generate error if calculated sound

          velocities are complex (can happen with bad CTD spikes)

    Jun  4, 2008:

        - improved [cruise.m] & [htmltable.m] to allow for spaces in the

          cruise ID (they are now replaced by underscores)

    Jun 26, 2008:

        - fixed bug in [begin_processing_step.m] (typo that broke

          'eval-expr' code)

        - changed default of p.interp_ctd_times = 1; in [loadctd.m]

          because this does not affect already regularly spaced ctd time

          series but improves time matching of irregularly spaced ones

    Jun 27, 2008:

        - changed version from 5beta to 5

    Jun 30, 2008:

        - made [geomag.m] portable for non-UNIX OSs

        - folded [geomag.m] into [mlocal/loadnav.m]

    Jul  2, 2008:

        - removed m & mlocal subdirs

        - BUG: geomag code in [loadnav.m] required additional "" to

               allow spaces in the path of the IGRF coefficient file

        - removed inappropriate warning from [loadctd.m]

IX_6:

    Jul 17, 2008:

        - changed [default.m] to version IX_6beta

        - BUG: start/end positions were adjusted incorrectly at the end

               of [loadctd.m] when a large time adjustment was required

               (mainly with elapsed time base in case the ADCP was

               started significantly before the CTD); significant

               changes required in [loadnav.m] [loadctd.m] and

               also a small (untested) change in [loadctdprof.m]

    Jul 27, 2008:

        - [loadctdprof.m] [loadnav.m]: nanmean() -> meannan()

    Sep 18, 2008:

        - BUG: p.navdata was not set when f.nav was set to ' ', causing

               [loadctd.m] to bomb. Solution: set p.navdata to 0 in

               [process_cast.m]; some modification in

               [loadctd.m] required as well

    Oct  7, 2008:

        - extensively modified ProcFig. 3 in [geterr.m] to detect

          inconsistent up/downlooker data; also changed colormap

    Oct 15, 2008:

        - replaced mean by median to get lat/lon in [loadnav.m]

    Jan  7, 2009:

        - implemented fix provided by Dan Torres to allow reading data

          files with extra data types; [loadrdi.m] marked with "%%%

          DT bug fix 2009-01-07"

        - tighten use of exist() from simple true/false tests to test

          for specific return values; affected files [default.m]

          [dragmodel.m] [getdpthi.m] [getinv.m]

          [getkzprof.m] [getserial.m] [htmltable.m]

          [htmltable2.m] [loadctd.m] [loadctdprof.m]

          [loadnav.m] [loadrdi.m] [loadsadcp.m]

          [prepinv.m] [process_cast.m] [rditype.m]

          [savearch.m] [saveprot.m]; this bug fix was

          implemented based on a problem report submitted by Anna Wahlin

          (who proposed a different fix)

        - finally removed laproc.m presolve.m resolve.m

    Jun 16, 2009:

        - bugfix related to short nav time series in [loadctd.m]

    Jul 22, 2009:

        - changed default of ps.smallfac in [default.m]

    Oct 27, 2009:

        - color-coded down/upcasts in ProcFig. 6 in [prepinv.m]     

    Oct 28, 2009:

        - modified p.ignore_beam for dual-head systems in [loadrdi.m]       

    Dec  1, 2009:

        - bugfix related to geomag date check in [loadnav.m]

    Jan 22, 2010:

        - adapted [loadnav.m] to Eric Firing's version of magdec utility        

    Mar  4, 2010:

        - updated documentation in [default.m]

        - BUG: removed shallow-profile special handling in [prepinv.m]

        - removed setting of pk.top in [default.m]

        - default sadcp_dtok to zero in [loadsadcp.m]

        - BUG: new ProcFig.3 did not work for downlooker-only systems; [geterr.m]

        - adapted [loadnav.m] to SOEST IGRF-11 magdec 

        - changed [default.m] to version IX_6beta2

        - emailed tgz file to Dan Torres

        - fixed minor bug in [prepinv.m]

    Mar  5, 2010:

        - changed [default.m] to version IX_6beta3

        - mailed new tgz file to Francois Ascani

        - changed all finite() to isfinite() WITHOUT UPDATING HEADERS!

        - changed [default.m] to version IX_6beta4

    Mar  8, 2010:

        - updated geomag/magdec.c to newest version

    Oct  1, 2010:

        - fixed bug in [checkbtrk.m] affecting profiles without valid BT

	Oct 18, 2010:

		- updated version in [default.m] to IX_6

		- pulled newest geomag version

		- cleaned up history for first Mercurial version

IX_7:		

	Jan  3, 2011:

		- changed IGRF11 validity to 2015 in [loadnav.m]

	Apr 14, 2011:

		- added TODO file with known bugs and potential improvements

	Apr 27, 2011:

		- updated version in [default.m] to IX_7

		- published on Mercurial server

IX_8beta:

	Jul 2011:

		- added new plot options to [geterr.m]

		- adapted [getinv.m] to new [geterr.m]

		- disabled bin-remapping in [loadrdi.m]

	Dec 30, 2011:

		- implemented user-provided bug fix in [fixcompass.m]

IX_8gamma:

	Jan  6, 2012:

		- removed lesqchol() from getinv.m because it does not

		  deal well with (near-)singular matrices; while this is

		  somewhat iffy in the sense that the new code can produce trash

		  where the old code produced some mix of nans and zeros;

		  however, if the data are really trash they will cause very

		  large velocities that are removed from the solutions, i.e. it

		  is hoped that this change does not confuse things

		- added imagnan.m

	Apr 26, 2012:

		- finally removed code that calculates diffusivity profiles

IX_8:

	May  4, 2012:

		- updated version in [default.m] [.hg/hgrc]

		- published on Mercurial server