Changes in version 1.32:

USER VISIBLE CHANGES

    o   Fixed vignette

    o   Added initial support to Generic Arrays

    o   Using BiocGenerics definition of normalize

Changes in version 1.24:

BUG FIXES

    o   Removed dependency on RConverters.h

Changes in version 1.22:

USER VISIBLE CHANGES

    o   Improved results by getProbeInfo();

    o   fitPLM, coefs and resids are now Defunct. Use fitProbeLevelModel,
	 coef and residuals respectively. 'coef' and 'residuals' follow the
	 standards used elsewhere in R;

BUG FIXES

    o   Fixed problem caused by the fact that oligoClasses had its own
	 annotation() method when BiocGenerics added a new one;

    o   Several fixes to probe selector, allowing 'target' to be used;

    o   PA Calls didn't know about target;

Changes in version 1.20:

USER VISIBLE CHANGES

    o   New getProbeInfo() function added, to simplify probe selection
	 without using SQL.

    o   New fitProbeLevelModel() function added. It allows probe-level
	 models ('plm' and 'medianpolish'), which can be used for QC.

    o   fitPLM, coefs and resids are now Deprecated. Use fitProbeLevelModel,
	 coef and residuals respectively. 'coef' and 'residuals' follow the
	 standards used elsewhere in R.

    o   Now using foreach for parallelization.

BUG FIXES

    o   Addressed issue in which probes without chr info would be removed
	 from pmChr, leading to results whose dimensions did not match PM
	 matrix (on TilingFeatureSet).

Changes in version 1.18:

USER VISIBLE CHANGES

    o   pmStrand() now available for Affy-Tiling arrays (by Kristof De
	 Beuf);

BUG FIXES

    o   Fixed message method when normalizing with large dataset support;

Changes in version 1.16:

USER VISIBLE CHANGES

    o   Using NEWS.Rd;

NEW FEATURES

    o   New summarization methods are available: PLM, PLMr, PLMrr and PLMrc.
	 These are implemented in the basicPLM function and also available
	 through the summarize() interface (method=c('plm', 'plmr', 'plmrr',
	 'plmrc'));

    o   fitPLM(), NUSE() and RLE() are now available;

    o   Presence/Absence calls are available through paCalls();

    o   MAplot method with advanced options;

BUG FIXES

    o   'pm<-', 'mm<-', 'bg<-' now work with ff_matrix replacements;

    o   boxplot() now uses the correct (default) ylim if transfo != log2;

    o   Fixed documentation for hist(), boxplot(), fitPLM();

    o   Fixed image for PLMset on NimbleGen arrays;

Changes in version 1.14:

USER VISIBLE CHANGES

    o   Added support to read gzipped XYS files directly.

    o   Added information for citation.

    o   Added getNetAffx method to retrieve NetAffx annotation for Exon ST
	 and Gene ST arrays.

    o   Reading in uncompressed XYS files is much faster than in 1.12.2.

    o   Added vignette for Exon/Gene ST arrays (rma and NetAffx annot)

BUG FIXES

    o   Fixed bug in boxplot / hist / MAplot

    o   Fixed issue that would appear with GeneFeatureSet objects, in which
	 it selections like obj[c(2, 2, 1),] would generate invalid
	 GeneFeatureSet objects

    o   Fixed bugs in image / bg / bgindex / bgSequence for Exon and Gene
	 arrays (due to changes in annotation)

    o   Added date/time of scan (for XYS/CEL files) to the protocolData
	 slot. For 1 channel applications, the field is called 'dates'; for
	 2 channel apps, there are 2 fields: 'dates1' and 'dates2'

NEW FEATURES

    o   New method, runDate(), to extract scan dates from FeatureSet objects

Changes in version 1.11:

USER VISIBLE CHANGES

    o   New vignette structure: Vignette 1: Overview (with multiple
	 examples) Vignette 2: Expression (NimbleGen data) Vignette 3: SNP
	 Genotyping (Affymetrix)

    o   Uses Imports rather than Depends. Users must explicitly load
	 Biobase, affyio, preprocessCore if they need functions from these
	 packages.

    o   Revised man page for rma.

NEW FEATURES

    o   Support to large datasets via ff package.  The user should load the
	 ff package in order to use this feature. Files used by ff are saved
	 to getwd(), unless the user changes the standard path.

    o   Support to parallel computing via snow package.  The user should
	 load *both* ff and snow in order to use the parallel features
	 implemented in the package (backgroundCorrect, normalize,
	 summarize, rma, normalizeToTarget). Cluster variable should be set
	 as options(cluster=makeCluster(...)).

    o   seqColors and darkColors, respectively, generate a vector of
	 sequential colors (in blue) and (qualitative) dark colors (based
	 onn Dark2 from RColorBrewer).

    o   plotM method for SnpSuperSet (plot genotype clusters).

    o   getAffinitySplineCoefficients estimates the affinity spline
	 coefficients.

    o   getBaseProfile estimates and optionally plots sequence effect like
	 gcrma.

BUG FIXES

    o   Fixed definition for probeNames() for Exon/Gene arrays.  (Mikhail
	 Pachkov)