0.0.1
  - New R package created.

0.0.2
  - FIX legend position in plotCorrelation.

0.0.3
  - Make read.codelink(raw=TRUE) default. Added changelog file.

0.0.4
  - Added require(genefilter) inside filter functions.
  - Split source file codelink.R into file.R, plot.R, filter.R, data.R and 
  class.R.

0.0.5
  - Added require(annotate) inside codelink.report().
  - FIX: Move codelink.report from plot.R to file.R.

0.0.6
  - read.codelink() now parses first file with scan(flush=TRUE,na.strings="") 
  to faster and detect valid number of genes in malformed txt exported files.  
  That allows the function to work now with different bioarray plattforms.
  - Changed format of changelog file.

0.0.7
  - Removed debug code from read.codelink().

0.0.8
  - read.codelink() set negative values to NA not 0.1 (as in limma).

0.0.9
  - renamed codelink.merge() to merge.codelink().
  - renamed codelink.report() to report.codelink().
  - Added documentation for merge.codelink() and report.codelink().

0.1.0
  - FIX enable of NA in plotMA() and plotDensities().
  - Added wrapper normalize.codelink() to normalize.loess() and 
  normalize.quantiles() from affy package.
  - read.codelink() set negative values to 0.01 This is for compatibility with 
  normalization functions from affy package, until new ones are build.
  - Added method write.codelink() to export data.
  - read.codelink() now can set decimal point (dec="." option).

0.1.1
  - Partially restored plotCorrelation() function, lost somewhere when source 
  files where splitted.
  - Added normalize.codelink.loess() a modified version of normalize.loess() 
  from affy package that allows NA values.
  - Added support for NA values in plotDensities() and plotMA().
  - Added neg.val option in read.codelink() to allow set negative values.  
  Default to NA.
  - Added method loess.affy in normalize.codelink() to allow normalization 
  using de default affy function.
  - Added method quantiles.affy in normalize.codelink() to allow normalization 
  using the default affy fucntion.
  - Modified method quantiles in normalize.codelink() to use 
  normalizeQuantiles() from package limma that can do quantiles normalization 
  with NA values.
  - Renamed plotDensities() to plotDensities.Codelink(), plotMA() to 
  plotMA.Codelink() and plotCorrelation() to plotCorrelation.Codelink() to 
  allow loading of packages like limma that have similar named functions to 
  other object types.
  - Renamed normalize.R to norm.R.
  - Changed some methods names in normalize.codelink().
  - Renamed normalize.codelink() to normalize.Codelink().
  - Make read.codelink() default neg.val=0.01.
  - Renamed read.codelink() to read.Codelink(), write.codelink() to 
  write.Codelink(), report.codelink() to report.Codelink() and merge.codelink() 
  to merge.Codelink().
  - Added "ALL" to type option in normalize.Codelink() to allow use all points 
  in normalization. Make it the default value.

0.2.0
  - Mayor changes to all functions. Change Codelink class description to 
  acomodate fields found in Codelink exported text data. Slots are named as the 
  corresponding fields. Some slots store information about different data 
  process. All functions changed have been named equal as the old ones plus 2 
  added. This is a transitional version. If all works those functions would be 
  renamed as the defaults in next version. A list of functions changed is:
  - Codelink class definition and other redefinitions: dim, as.matrix, [, etc.
  - read.Codelink2() new slots assignment. Reads Spot_mean and Bkgd_median.
  - merge.Codelink2() support to new Codelink class. Some improvements.
  - bkgdCorrect.Codelink() bakground correction based in limma ones. Supported 
  methods: none, subtract, half, edwards and normexp (not working always).
  - normalize.Codelink() adapted to new Codelink class.
  - plotMA.Codelink2() adapted to Codelink class. Detects automatically slots.
  - plotDensities.Codelink2() as above.
  - plotCorrelation.Codelink() finished implementation. As above.
  - write.Codelink() support to new Codelink class. FIX for header information, 
  in progress.
  - report.Codelink2() initial support to new Codelink class.
  - Removed assignment of cv slot from merge.Codelink(). Not sure yet how to 
  make it.

0.2.1
  - Support for autodetecting decimal point type.
  - FIXED write.Codelink() header and allow to specify decimal type.

0.2.2
  - Improved as.matrix() method.
  - Support for reading Raw_intensity and Normalized_intensity with 
  read.Codelink2()

0.2.3
  - Renamed all new functions to the default ones: read.Codelink2() to 
  read.Codelink() etc.
  - Removed all old functions except report.Codelink() not yet updated.

0.2.4
  - FIXED report.Codelink2() and moved to report.Codelink(). All function 
  deleted.
  - normalize.Codelink() allow return of log2 values. Info stored in 
  'Log_transformed' slot in Codelink object.
  - changed plot functions to detect 'Log_transformed' slot.
  - make 'quantiles' default normalization method for normalize.Codelink().

0.2.5
  - Added support to cv calculation in merge.Codelink() to later filter of 
  genes.
  - FIXED class.R to add support for 'Normalized_intensity' and 
  'Normalized_intensity_cv' slots in '[' method.
  - merge.Codelink() stop if 'names' argument is NULL.
  - merge.Codelink() allow for log2 the 'Normalized_intensity' slot. That is 
  because merging is not the same on log or not log data (Arithmetic mean vs 
  Geometric mean).
  - normalize.Codelink() code clean up.

0.2.6
  - Header information in exported file with export.Codelink() added.
  - Default values for 'BkgdCorrection_method', 'Normalization_method' and 
  'Merge_method' slots changed from 'NA' to 'NONE'.

0.3.0
  - Automatic Sample name detection from header (restored).
  - Minor modifications to header information in write.Codelink().
  - Some code clean up in all files.
  - Moved to colorRampPalette() in plots.

0.3.1
  - log2.Codelink() function to apply log2 to a Codelink object.
  - Support for highlighted points in plotMA.Codelink().
  - Clean up console output from read.Codelink() plus summary info added.
  - read.Codelink() read 'Bkgd_stdev' slot when available and if 'Spot_mean' is 
  read.
  - snr.Codelink() calculate Codelink SNR as described.
  - SNR() calculates SNR in read.Codelink()
  - plotMA.Codelink() support labeling by 'SNR' slot.
  - FIX: NAs handle in plotMA.Codelink() (Now labels are correctly displayed 
  using 'Spot_mean')
  - Support for 'Bkgd_stdev' and 'SNR' slots in class definitions.
  - Support for 'SNR' in merge.Codelink().
  - FIX: Some weird things in merge.Codelink()
  - fc.Codelink() method added to select genes based in fold change.

0.3.2
  - FIX: plotMA.Codelink() correct position of legend in "Probe_type" labeling.
  - plotMA.Codelink() support 'plot.theme' option to better visualization on 
  screens (i.e. 'black theme')
  - report.Codelink() accepts a list of genes (character vector) again.
  - Added support to merge.Codelink() for use of 'Raw_intensity' slot from 
  unnormalized data.
  - Renamed 'Normalized_intensity_cv' slot to 'Intensity_cv' as 
  merge.Codelink() can be applied to unnormalized data.

0.3.3
  - Detection of product type from file. An error message is produced of 
  different arrays are loaded together.
  - FIX some information log in read.Codelink() and merge.Codelink().

0.4.0
  - Modified plotDensities.Codelink() to better visualization.
  - Remove Bkgd_stdev slot after computing SNR in read.Codelink().
  - Added SNR label in plotCorrelation.Codelink()
  - FIX: SNR label in plotMA.Codelink() and plotCorrelation().
  - FIX: Some fixes in plotCorrelation.Codelink().

0.4.1
  - Added Batch_name slot prompt in read.Codelink() to use as Batch Name for 
  analysis.
  - Added support for Batch_name slot in plotMA.Codelink(), 
  plotDensities.Codelink() and plotCorrelation.Codelink().
  - Initial support for Codelink Flags on file loading: MSR and I flagged spots 
  are set to NA.
  - Change in normalize.Codelink() from normalize.quantiles in package affy to 
  normalizeQuantiles from package limma that allows for NA.
  - Minor FIX in Normalize_method slot identifier (Only method not package 
  used).
  - Removed high.list support in plotMA.Codelink() as it is not working as 
  expected. To be fixed later...
  - Removed filter.R file with filter function for genefilter package (Not 
  used).

0.4.2
  - Added support for all Quality_flag counts. 'M', 'I' and 'X' spot are set to 
  NA.
  - printHead() in class.R renamed to printHead.Codelink() to compatibility 
  with limma.
  - Automatic detection of Report name in Codelink files, assigned to 
  'Batch_name' slot.
  - Smart positioning of legend in plotting functions. Also remove some 
  unnecessary code.
  - Restoring highlighted genes support in plotMA.Codelink(), high.list is a 
  logic vector indicating which genes of the object passed to highlight.

0.4.3
  - Added option legend.x on plotMA.Codelink() to allow change position of 
  legend.
  - Added plotCV.Codelink() for density plot of C.V. of merged objects.
  - Added cutCV.Codelink() to calculate cutoff for C.V. of merged objects.
  - Use normalize.loess.Codelink() with method 'loess' in normalize.Codelink() 
  to allow for NAs.
  - Allow change title in plotMA.Codelink(), plotCorrelation.Codelink() and 
  plotDensities.Codelink(). Default tells measurement used (Spot_mean, 
  Raw_intensity or Normalized_intensity).
  - Read Product, Project and Experiment information from data files into data 
  object as Product_name, Project_name and Experiment_name.
  - Change Experiment_name as default information title in plot functions.
  - 'C' flagged spots (Background contaminated) are no set bye default to NA.
  - Allow custom intensity set based on flags. read.Codelink() accepts option 
  flag that is a list with default (M=NA,I=NA,C=NA,X=NA).

0.4.4
  - Added subset support in plotCV.Codelink().
  - Added cex for legend in density plotting functions.
  - FIX some bugs in report.Codelink() to handle multiple Genbank accession 
  numbers.
  - Some minor FIXES in plotMA.Codelink().

0.4.5
  - high.list in plotMA.Codelink() need to be numerical vector indicating index of Codelink object to be plotted.
  - better handling of SNR calculation in read.Codelink().

0.4.6
  - FIX warnings about documentation in R CMD check. Renamed merge.Codelink() 
  to mergeCodelink().
  - Removed main option from plotDensities.Codelink().
  - Moved doc directory to inst directory as in other packages. Now doc will
    contain the vignette of the package.
  - Changed license to GPL.

0.4.7
  - Added other, other.names and ordered arguments to report.Codelink() to add
	extra column with names (thinking on fold changes) and to be ordered on 
	that column.
  - FIX genes.type error in report.Codelink().
  - na2false() function set NAs to FALSE in logical vectors. Used in
    fc.Codelink().
  - Added na2false() before return in fc.Codelink().
  - Added selCV.Codelink() to select genes based on CV cutoff.
  - Renamed mergeCodelink() to mergeArray.Codelink() (maybe not the last
    time :) ).

0.4.8
  - Added arbitrary column addition to report.Codelink() passed as list. The
    argument 'ordered' is renamed to 'other.ord' and accept the slot name or
    the slot number of the list to order genes by. The argument other.names is
    removed as names are obtained from the list.

0.4.9
  - Minor FIXES to plotting functions.
  - Added codelink.example to data directory. Contain an example object for
    Codelink class.
  - Added first version of a vignette.
  - Removed methods not actually used in bkgdCorrect.Codelink() until they
    become well implemented.
  - Intensity_cv slot renamed to CV.
  - FIX File_name and Sample_name slots subsetting in '[' method.

0.5.0
  - FIX arguments passed to htmlpage() in report.Codelink().
  - Some changes to plotting functions.

0.5.1
  - Project_name and Experiment_name are not stored now (as they could be very
    variable) but yet parsed by read.Codelink().
  - Removed dependence on Experiment_name and Project_name on plotting
    functions.
  - Rewrite of plotDensities.Codelink() as in plotCV.Codelink(), array
    argument changed to subset.
  - Improved codelink.example.
  - Improved vignette.

0.5.2
  - Support missing Sample Name in Codelink files. If no sample name supplied
    it is used the file name. Yet not support arbitrary long headers.

0.6.0
  - Mayor rework of read.Codelink(). Now allow different header files but of
    the same product type and chip size. New functions readHeader.Codelink(),
    dec.Codelink() and chipSize.Codelink().
  - FIXED some bugs in flag dependent assignment (missing flags).

0.7.0
  - Version pump to enter BioC devel repository.
  - Added biocViews slot to DESCRIPTION file.
  - Some output polish of read.Codelink().
  - FIX bug in determining if Bkgd_stdev column exists.
  - Added 'preserve' argument to read.Codelink() that preserve Bkgd_stdev
    slot.
  - Added 'verbose' argument to read.Codelink() to set the level of
    informative output.
  - FIX use of argument 'dec' on read.Codelink().
  - Some changes in class definitions: LargeDataObject-class removed as used only in
    Codelink-class. Method 'show' is now a Codelink-class method. Added some
    basic documentation of Codelink-class.

0.7.1
  - Added NAMESPACE to the package.
  - Mayor API change: Almost all functions have lost the '.Codelink' tail.
    Some functions have been totally renamed. See vignette for detailed
    information about the new functions. Remain with '.Codelink' real S3
    methods dim.Codelink() and as.matrix.Codelink().
  - Changed Depends: R (>= 2.2), limma, annotate in DESCRIPTION file.

0.7.2
  - Added createWeights() function to create weights based on Probe_type.
  - Remove unused argument 'type' from normalize() function.
  - Updated vignette.
  - Renamed group argument in mergeArrays to class.
  - Added some more functions to NAMESPACE (logCodelink, na2false,
    createWeights).

0.7.3
  - Mayor API change: Simplified slot names for Codelink object. See vignette
    for a detailed description of new names and correspondence with old names.
  - Added preserve argument to bkgdCorrect() and normalize() to allow
    preservation of Smean, Bmedian and Ri slots.
  - readCodelink() don't set spots flagged 'X' as NA by default.
  - Updated vignette.

0.7.4
  - Changed 'file' argument in readCodelink() to
    'file=list.files(pattern="TXT")' which is the most common used as Codelink
    files have extension TXT.
  - Better handling of 'Product' detection. If missing, product slot is set
    to unknown and a warning is shown. If two product description disagree and
    error is shown and the reading of files stopped.

0.7.5
  - Support for multiple flags in the same probe. The minor value or NA if
    found is used.

0.7.6
  - FIX normalize() to warn about not background corrected data (it didn't
    normalize but also it didn't say nothing about it).
  - Improved snr label in plotMA(); now three levels are shown.
  - merge.snr option to allow not to mean SNR in mergeArray(). Default: TRUE.
  - Removed fc() in favour of fc2Cond() a more intuitive interface for fold
    change computation.

0.7.7
  - Added snr.cutoff option for cutoff definition (default=1) used in
    plotMA() to label spots by SNR.

0.7.8
  - Changed lowess line parameters in plotMA() for better visualization.
  - FIX lowess line computation.

0.7.9
  - Added argument 'subset' to plotMA() allowing plotting only a subset of 
  spots,  based in 'type' slot.
  - Added argument 'pch' to plotMA() to control pch used to general points.  
  May be useful when plotting subsets of spots.

0.7.10
  - Check for files detected in readCodelink().
  - Check for correct header file in readHeader().
  - Speed up of readHeader().

0.7.11
  - Added imageCodelink() function to plot arrays yet experimental.
  - FIX flag based assignment in readCodelink().
  - FIX check for log intensities in normalize().
  - FIX log Ri in normalize().

0.7.12
  - Rewrite of writeCodelink().
  - FIX htmlpage link in man pages.

0.7.13
  - FIX a bug in writeCodelink().
  - Rename temporally plotMA() to plotMA2() as mask plotMA() from limma
  package. The plotting of MArrayLM object is lost.
  - This version is not committed to svn repository.

0.99.0
  - Renamed again plotMA2() to plotMA().
  - Added support for "MArrayLM" object from limma package in plotMA().
  - Version set to 0.99.0 to be (hopefully) part of BioC 1.8.
  - Some FIXES in package vignette.

1.0.0
  - Version changed to BioC 1.8 release.

## BioC 1.9

1.1.0
  - Version changed for BioC 1.9 devel.

1.1.1
  - Added support for normexp background adjustment from limma package.
  - Added offset argument to bkgdCorrect() with default=0.
  - Updated maintainer information.

1.1.2
  - Improvements to imageCodelink().
  - New function arrayNew() creates calls x11() device to set suitable
  dimensions to be used with imageCodelink().

1.1.3
  - Change order of parameters in arrayNew().
  - Fix setting of slot method$background when using normexp method.

1.1.4
  - Small FIX in documentation.
  - FIX Codelink-class.Rd as yet had some affy packages information...
  - Added method "median" to normalize().

1.1.5
  - Added experimental support for XLS exported files with readHeaderXLS() function.
  - Set parameters fill=TRUE and blank.lines.skip=TRUE in decDetect() to make it more robust.
  - Added methods to dependences in DESCRIPTION file. Added a check for methods load in .onLoad() function.

1.1.6
  - FIXED a bug introduced in readCodelink() that prevented reading of files.

1.2.0
  - Version number increased for BioC-1.9 release.

1.2.1
  -  FIX read of files without "Logical_row" and "Logical_col" fields.

## BioC 2.0

1.3.0
  - Version number increased for BioC-2.0 development.

1.3.1
  - Some code cleanup in readCodelink().
  - Remove row.names=1 from readCodelink() to allow files without "Idx" field to be read.
  - FIX read of files without "Logical_row" and "Logical_col" fields.
  - Small change in changelog.txt to reflect BioC releases.

1.3.2
  - Experimental: Added generic function for Codelink objects to 
  plotDensities().
  - Changed example data 'codelink.example' to resolve issues with the show 
  method, creating a vignette of several hundreds of pages.
  - Removed unnecessary line 'Packaged' from DESCRIPTION as it is automatically 
  generated by the build system and remains from my initial steps with R 
  packages...
  - Removed also dependency on R version as I really don't know which minimum 
  version is needed. As I always use the current R version, this is the 
  recommended one.

1.3.3
  - FIX as.matrix.Codelink for R-2.5.x.
  - Reorganization of some code in class.R
  - class.R renamed to Codelink-class.R

1.3.4
  - Added argument 'check' (default=TRUE) to readCodelink() that allows 
  checking for column order consistency. Some versions of the Codelink software 
  change the probes order in the exported file. You get weird results if that 
  is the case, but may not notice the cause.
  - Added argument 'fix' (default=FALSE) to readCodelink() to allow a fix of 
  the probe order consistency problem. Probes are ordered based on Feature_id 
  if present, which is an unique identifier based on probe position.  If not a 
  dirty fix is tried using Probe_name, which may end messing a little bit some
  duplicated probes, like control and fiducial as well as a few discovery 
  probes spotted in duplicate.
  - Improvements done to the vignette including new entry for imageCodelink() 
  and for the new readCodelink() arguments.
  - Added CITATION file into inst directory.

1.3.5
  - Added Collate field to DESCRIPTION to control the order .R files are read.  
  Added import(method) to NAMESPACE.

1.3.6
  - Changed use of LOCUSID to ENTREZID in reportCodelink().

1.3.7
  - Updated example data codelink.example to new S4 representation in R-2.5.
  - Added SaveImage: yes in the DESCRIPTION file.

#Released BioC-2.0: 2007-04-26

1.4.0
  - Version number increased for BioC release.

1.5.0
  - Version number increased for next development version.

1.5.1
  - Removed SaveImage in DESCRIPTION file as it is Deprecated in favor of 
  LazyLoad: yes in R-devel.

1.5.2
  - FIXED codelink.example.Rd (unmatched braces) and small improvements in 
  documentation, following 'Writing R Extensions' guidelines.
  - FIXED changelog.txt file and improvements.

1.5.3
  - Some minor fine tunning of plotMA(). Probably this needs major re-work.
  - Changed default pch value for high.list to 21 with col="gray" and 
  bg="orange" in plotMA().
  - Updated CITATION file.
  - Re-write of plotMA, clean up code and more smart arguments. Improved 
  documentation.

1.5.4
  - FIXED: no info stored when using normalize() with method = 'median'.
  - FIXED: header was not writen to exported file in writeCodelink().
  - Added SNR to output of writeCodelink(). Improved header information.

1.5.5
  - plotMA() now plot mean SNR values when label = "snr" instead of individual 
  pairs. In addition, it can use SNRmean if present in a MArrayLM object (which 
  has to be added before by hand) or it can be passed to the snr argument. This 
  allows showing the SNR information with fitted objects.
  - Added weights support for method CyclicLoess ('loess' method) in 
  normalize() through the corresponding argument.
  - Added verbose option to normalize(), default to FALSE. Set same option in 
  bkgdCorrect() to FALSE too.

1.5.6
  - Some improvements on the vignette as well as adding information about the 
  new methods and options.
  - Code cleanup in file.R. Now readCodelink() try to read the Feature_id field 
  if present. This will allow in the future the transition from Codelink-class 
  to CodelinkSet-class based object (class derived from eSet). The Feature_id 
  is stored in the 'id' slot.
  - Added initial files CodelinkSet-class.R CodelinkSet-methods.R and 
  CodelinkSet-tools.R with definition of CodelinkSet, a class based on eSet.  
  The function readCodelink2() readsdata into the old class and converts it to 
  a CodelinkSet-object using the function Codelink2eSet(). Some methods are 
  available but most the functionality is not working. This is a very first 
  step for moving to eSet-based objects. The package now depends on Biobase.

1.5.7
  - Modified getInfo() to retrieve all preprocessing info from experimentData 
  if no argument what is suplied.
  - Added codPlot() an experimental general interface for all plotting 
  functions. Implemented codPlotMA() function.
  - Added column meanSNR to phenoData object. This way, it don't have to be 
  computed every time we want to plot SNR in a codPlotMA. Added method 
  meanSNR() to extract this information.
  - Implemented normalization interface codNormalize(). Implemented 
  pre-processing interface codPreprocess() that includes codCorrect() and 
  codNormalize().
  - Added functions normalize.median() and bgcorrect.normexp() to manage common 
  computing.
  - Implemented codPlotMA() for CodelinkSet and MArrayLM objects. Implemented 
  codPlotDensities() for CodelinkSet objects. Added basic function plotma() for 
  plotting common stuff in codPlotMA().

1.5.8
  - FIXED: codPlotMA() for MArrayLM-class. Added method for codPlot().
  - When trying to correct or normalized data already corrected or normalized 
  the functions show a warning and return the original data object.
  - FIXED: plotma() in managing presence of type/snr information.
  - ADDED: support for highlight spots in codPlotMA().

1.5.9
  - Added left features to codPlotMA() method.
  - Added partial argument support to 'what' argument in codPlot() by using 
  match.arg() function.
  - Implemented codPlotImage() function. Added support function 
  getChipDimensions() to get the grid/subgrid values for the knwon chips.
  - Modified arrayNew() to use getChipDimensions(). Added function chipDevice() 
  that creates an X11 device with the correct dimensions, calling arrayNew().

1.5.10
  - FIXED: islog is only assigned for Codelink-objects in plotMA().
  - CHANGED: default colors for highlighted probes in plotMA() and codPlotMA() 
  are set to cyan for background and blue for borders.

1.5.11
  - Imported especific variables from packages in Depends. In particular,   
  normalizeQuantiles (as normalize.quantiles) from package limma and lookUp and 
  htmlpage from package annotate.
  - ADDED: If non log2 intensities are passed to codPlotDensity() then the 
  transformation is done. Now, ... is passed to lines() and lwd is a parameter 
  with default 2.
  - ADDED: Interface to codPlot() for old Codelink-objects. It will be 
  accessible until Codelink-class is deprecated.

1.5.12
  - FIXED: correct subsetting of logical$row and logical$col in Codelink 
  objects.
  - ADDED: functions to filter based on SNR, filtersnr is a function for use 
  with the genefilter package. filterSNR() are wrapper functions for Codelink 
  and CodelinkSet objects.
  - ADDED: codUniverse() creates a universe of the especified annotation ids to 
  be used with GOstats. Currently supported and tested ENTREZID. Works on 
  Codelink and CodelinkSet objects.

1.5.13
  - ADDED: functions writeCluster() to output intensity values suitable for use 
  with Cluster/TreeView. Currently methods are for Codelink, CodelinkSet and 
  MArrayLM objects. The name is likely to change in favor of the new API 
  convention, probably as an all-purpose exporting function.

1.5.14
  - FIXED reportCodelink() when NAs appeared in HTML, changed to blank.
  - MODIFIED writeCluster() format of annotation and fixed NA in some 
  annotations.

1.5.15
  - FIXED createUniverse() using always rwgcodGO as GO environment. Now it uses 
  the one determined by the chip.
  - FIXED documentation of as.matrix() and dim() where it was using the full 
  names of the S3 methods as.matrix.Codelink() and as.dim.Codelink().
  - FIXED encoding issue with arraySize.Rd. Small improvements.
  - ADDED \dontrun{} to the pertinent \examples{} section in the documentation 
  as is the proper method to display not run commands. Some minor improvements 
  in the documentation.

1.5.16
  - FIXED weird return in codCorrect().
  - ADDED experimental C code in src/ to fast reading of Codelink files. The 
  function readCodelinkFiles() works pretty much the same as readCodelink() but 
  is much faster. Currectly it reads data as a Codelink object. In addition, it 
  reads also the feature_id field, that can be used for creating suitable 
  CodelinkSet objects.
  - ADDED this comments for the first time to the svn log.
  - ADDED summary method for Codelink objects. Actually it outputs a summary of 
  the flags.

1.5.17
  - FIXED improper use of '&' instead of '&&' inside 'if' statements (suggested 
  by Seth Falcon).
  - REMOVED C code for fast reading of codelink files. The function strsep() is 
  not portable to windows hence the inclusion of this code must wait until a 
  different approach is found.

1.5.18
  - FIXED flag application in readCodelink() so that it doesn't reset all 
  values when setting one. Simplified flag based filter.
  - ADDED summaryFlag() method for Codelink objects. Included in the vignette.
  - FIXED some improper use of '|' instead of '||' too (forgot last time).

1.6.0
  - Bioconductor 2.1 release.

1.7.0
  - Development branch created.

1.7.1
1.7.2
  - Added method annotation() to extract that information when available.  The 
  annotation information is now guessed for new objects and stored in 
  $method$annotation.- Changed output format for writeCluster(), it is now use 
  the gene symbol for annotation.
  - FIXED typo in codelink vignette: mergeArray() doesn't have a 'group' 
  argument, changed to class.
  - FIXED warning about functions lacking description field in documentation.
  - From now on the changelog will be appended to the svn commit message.

1.7.3
  - Experimental class CodelinkSet is now derived from ExpressionSet rather 
  than eSet, allowing some nice integration with other pakcages (like limma).

1.7.4
  - Added method for ExpressionSet to writeCluster().
  - Revert previous change from exprSet to ExpressionSet in the vignette and
  mention current work in an ExpressionSet derived class.
  - FIXED direct access to slots in codCorrect(), now it uses ancestor
  methods (i.e. robust to API changes).

1.7.5
  - added new function c2e to better convert Codelink-objects to CodelinkSet.
  - c2e accepts a phenoData object (like those read by
  read.AnnotatedDataFrame().
1.7.6
  - added phenodata and feature data arguments to readCodelink2().

1.8.0
  - version for bioc-2.2.

1.9.0
  - version number for bioc-devel.

1.9.1
  - import 'annotation' and 'exprs' from Biobase to avoid warnings in R-2.8
  (devel).

1.9.2
  - added vignette describing the CodelinkSet class.
  - fixed notes about removing empty entries in .Rd files.

1.9.3
  - fixed high intensity probes turned NA by loess normalization.
  - code clean up in writeCluster().

1.9.4
  - small fixes to vignette.
  - added small section for exporting data to vignette.
  - writeCodelink() now outputs ACCESSION and ENTREZID values.
  - added method writeCodelink() to CodelinkSet.

1.9.5
  - added writeCodelink() example to vignette.
  - new wrapper function readCodelinkSet().
  - fixed writecluster symbol retrieval problem (now using lookUp from
  annotate package).

1.13.4
  - Move changes to NEWS file.