Please note that the release history presented here only goes back as far as version 5.
Version 10.10.0.0 - Thursday, May 9, 2024
Fix for discretization editor in 'Add Nodes From Data' in User Interface.
Fix AllEvidence log-likelihood test in RelevanceTree inference engine.
New tutorials.
Version 10.9.0.0 - Tuesday, February 21, 2023
Fixed certificate issue with Windows Desktop App Installer.
Fixed 'Attempting to JIT compile methods... in aot-only mode' on macOS M1 chips.
Expressions now support accessing Links, LinksIn and LinksOut from nodes.
Version 10.8.0.0 - Thursday, December 8, 2022
Fixed corrupt file on save in native desktop versions.
Version 10.7.0.0 - Thursday, December 1, 2022
Fixed 'negative width or height' error with query explorer
Fixed 'Trial/Activated' refresh issue on Start Page, after license entered/cleared
Fixed optimization error message when no design variables available, when function nodes are present.
Version 10.6.0.0 - Saturday, November 26, 2022
API. New query lifecycle interface added to inference algorithms, to allow dynamic query options.
Effects analysis now supports Disjunctive Cause causal AI algorithm
Causal optimization now supports Disjunctive Cause algorithm.
Disjunctive Cause algorithm now supports zero interventions.
Fixed 'status bar not updating' issue on MacOS
Fixed 'initial clicks not responding' issue on Desktop Browser versions.
Fixed 'None' node format Argument exception in User Interface
Fixed slow states editor, for larger number of states
Fixed blank query explorer on startup and theme change
Fixed 'Overflowing multi select controls' in User Interface
Version 10.5.0.0 Beta - Monday, November 21, 2022
Export/Load link constraints in structural learning User Interface.
Batch query User Interface converted to Wizard.
Parameter learning distributions can now be selected by group.
Fixed 'Argument Null Exception' error when saving a data map in User Interface.
Version 10.4.0.0 Beta - Sunday, November 13, 2022
Mac desktop app with installer, in addition to existing browser desktop version.
Windows desktop app with installer, in addition to existing browser desktop version.
New example networks added.
Many minor issues fixed since last preview.
Version 10.3.0.0 Beta - Saturday, August 6, 2022
Distribution editor V-Order support.
Additional Node formatting options now available.
Default Likelihood sample updated to 10k.
Custom D-Separation query added to cross platform User Interface.
Shortcut key support added to cross platform User Interface.
Node stats bar not supports configuration in settings.
Node stats bar now supported in pdf/clustered column/frequency column displays.
Version 10.2.0.0 Beta - Sunday, June 19, 2022
New cross platform User Interface.
New dark theme.
Causal effects analysis.
Causal inference with Direct effects (in addition to existing support for Total effects)
Causal inference with Unobserved confounders
Advanced causal query editor, supporting multiple interventions and treatments.
FrontDoor criterion/causal inference algorithm.
Backdoor criterion/causal inference algorithm.
Disjunctive Cause criterion/causal inference algorithm.
Abduction helper added to the User Interface.
Display Mean +- Stdev on discretized nodes in User Interface.
Causal Backdoor graph can be generated.
Causal Indirect graph can be generated.
Temporary data source can quicly be created by sampling the current network.
Optimizer/causal optimizer continuous bounds can be saved/loaded in the User Interface.
Auto Insight tool now supports interventions.
Default random number generator now consistent across different APIs.
Table and function expressions now support random number generation.
Sum Absolute Error added to regression statistics in the API.
Mesh query now supports interventions
Mesh query now supports both 2D and 3D plots
Data explorer now supports interventions
Mesh query now supports functions on z axis
Data Count API has been deprecated
Change in Noisy Order on a link now only effects a single child distribution
Version 9.5.0.0 - Thursday, April 15, 2021
Fixed unrecoverable error in user interface when switching to noisy nodes.
Removed causal options, as not currently used.
Intervention evidence can now be set using SHIFT click in Gaussian charts.
Added support for noisy experience tables.
Discrete nodes (in probability bar display mode) show an ellipsis row when any states are hidden.
About box now displays current license key (if any).
Version 9.4.0.0 - Saturday, January 16, 2021
Optimization now supports target value, in addition to maximize and minimize.
Intervention evidence can be set using SHIFT click in network viewer.
Causal optimization options now available as column in optimization wizard.
Fixed setting continuous evidence in auto-insight analysis tool.
Error message improved when indexer syntax attempted in expressions.
Version 9.3.0.0 - Saturday, November 7, 2020
Optimization framework: search sets of evidence to minimize or maximize a function, continuous variable or discrete state.
Genetic optimization algorithm to optimize evidence.
Genetic simplification algorithm to simplify optimized evidence.
Joint queries (incl. Decision graphs): mean values (incl. utility values) now return NaN if their associated probability is zero.
Fixed max propagation with function nodes.
New switch template added to expression editor templates.
Network chart visuals improved.
Standard deviation, variance and means added to Gaussian chart visuals.
Version 9.2.0.0 - Thursday, August 6, 2020
Improved error message when an expression invocation does not include the method name.
Max propagation with function nodes fixed.
Error when invalid syntax in expressions improved.
Expression Code Item missing header fixed.
Version 9.1.0.0 - Wednesday, August 5, 2020
Corrupt expression evaluation in Table expressions and Function nodes fixed.
Version 9.0.0.0 - Monday, August 3, 2020
Table expressions. Use code to define the parameters of tables, using stats, variables, math, and other complex expressions. Multi-statement expressions supported.
Function nodes. Augment networks with function nodes, that evaluate complex expressions based on queries, log-likelihood, other functions, etc....
Interventions. Discrete, continuous & temporal
Frequency Density charts within nodes in the network viewer.
Clustered column charts within nodes in the network viewer.
Probability Density charts within nodes in the network viewer..
More scalable alternative decision graph algorithm (SPULight).
Per variable numeric formatting in network viewer.
Per variable min/max chart formatting in network viewer.
BREAKING CHANGE: DecisionAlgorithm.Default changed to SPU Light, and older SPU algorithm name now has Full suffix.
DEPRECATED: Java 7 support removed. 8+ now supported.
DEPRECATED: JavaScipt API deprecated, in favor of web assembly alternatives.
Option to auto-query network on open
Toolbar button to display node/variable descriptions.
API: Topological sort with depth.
Batch query export to Excel or Csv.
Data sampling export to Excel or Csv.
Version 8.20.0.0 - Wednesday, July 1, 2020
Value of information updated to support zero probability hypothesis states.
Decision graph algorithm fix. Algorithm was not ordering decisions in certain siuations which increased the chance of less optimal results.
Version 8.19.0.0 - Sunday, March 15, 2020
QueryOptions.InconsistentEvidenceMode allows control of when InconsistentEvidenceException is raised. Available in both the API and as an option in the User Interface.
Option added to the User Interface to always calculate the log-likelihood and report it in the status bar.
Version 8.18.0.0 - Saturday, February 15, 2020
Fixed online learning (adaptation) with deterministic distributions.
Pattern analysis documentation updated.
Version 8.17.0.0 - Saturday, January 4, 2020
Tree query: tree width now supports continuous variables.
Fixed initialization bug in Parameter Learning when temporal latent node is a child of a non temporal latent node
Version 8.16.0.0 - Thursday, November 14, 2019
API now supports parameter learning with soft/virtual evidence.
Version 8.15.0.0 - Saturday, November 2, 2019
Improvements to final solutions reached by both Clustering structural learning and Hierarchical structural learning algorithms.
Version 8.14.0.0 - Tuesday, October 29, 2019
Important improvements to cluster count estimation, used in both Clustering structural learning and Hierarchical structural learning.
Version 8.13.0.0 - Saturday, October 26, 2019
Important improvements to parameter learning initialization, leading to improved solutions.
Improvments to batch query, data sampling, and other windows on high DPI displays.
Reduced default parameter learning tolerance with log-likelihood convergence, made possible due to improvements in the initaialization routines.
Version 8.12.0.0 - Saturday, September 7, 2019
New distribution editor, with filtering, sorting, grouping and dynamic validation.
Parameter count now uses a double precision floating point, to avoid overflow.
Support added for default partition columns in the User Interface.
Structural learning algorithms now throw an exception if data is mapped to nodes not being learned.
DateTime columns are automatically deselected in 'Add nodes from data' wizard
Add node menu item in Network editor fixed.
Fixed Data explorer map to interval error when 'StateNotFound' is missing
Version 8.11.0.0 - Wednesday, July 24, 2019
Dynamic D-Separation display.
Complex D-Separation query.
Version 8.10.0.0 - Wednesday, July 17, 2019
Fixed regression in 8.9 - online learning (adaptation) with noisy nodes.
Version 8.9.0.0 - Monday, July 15, 2019
Fixed online learning (adaptation) with noisy nodes.
Search & score structural learning now supports latent variables
Version 8.8.0.0 - Saturday, April 13, 2019
Fixed add 'Custom Query' to 'Query explorer'.
Fixed missing Java docs for classes implementing interfaces.
Improved noisy node documentation
Version 8.7.0.0 - Saturday, March 2, 2019
Fixed in-place discrete temporal node display with large time range, which was showing as dimmed or all white.
Update: Add nodes from data now supports grid layout and adding nodes in compact view mode.
Version 8.6.0.0 - Wednesday, January 30, 2019
New in API: D-separation calculator.
New in API: Empirical density now includes inverse Cdf. (e.g. for calculating confidence intervals from network samples).
New: VariableDefinition (add nodes from data) now supports EmptyStringAction.
Fix: Default add node button now correctly displays manual add node dialog, not add nodes from data dialog.
Version 8.5.0.0 - Tuesday, January 1, 2019
Fix: Network.Load fixed for Mono targets such as Blazor and Unity.
Version 8.4.0.0 - Saturday, December 15, 2018
New: Clustering structural learning algorithm, automatically defines a latent variable for a model. Can be used in conjunction with other learning algorithms.
Fix: Now window maximize/minimize buttons do not disappear on certain OS
New: Improved .NET API documentation layout and search.
New: Added Munin sample network.
New: Multiple Information column properties can be changed at once.
Fix: Confusion matrix grid and errors now display properly when not strings.
New: Partitions and MaximumIterations options added to Hierarchical clustering algorithm.
Version 8.3.0.0 - Saturday, October 6, 2018
Fix: 'Unable to load one or more of the requested types. Retrieve the LoaderExceptions property for more information' error was raised on .NET version 4.6.1 and 4.6.2
Version 8.2.0.0 - Monday, October 1, 2018
Auto-insight and pattern analysis now support continuous targets
Auto-insight user interface improvements
Pattern analysis user interface improvements
Auto insight and pattern analysis now support distance measures between 0-1 for easier interpretation
API: Jensen Shannon Divergence
API: Table.set(...) with varArgs now correctly checks for an owner to stop edits when assigned to a node.
API: Java version now correctly manages changes made to a network when re-using the same inference engine
API: Evidence.Clear now correctly resets the weight to 1 when no evidence is present.
Version 8.1.0.0 - Sunday, September 16, 2018
Pattern analysis helps differentiate a discrete target state from the other states in the same variable. Useful for understanding clusters in a cluster/mixture model.
Fix for Java 9/10 deprecation/removal of javax.xml.bind.* (Java breaking change)
Data mapping editor updated
BREAKING CHANGE: .NET 3.5 no longer supported
BREAKING CHANGE: Java 6 no longer supported
BREAKING CHANGE: Entropy class removed from Analysis namespace. Use more advanced version in Statistics namespace.
BREAKING CHANGE: Cross validation classes moved to Data namespace
BREAKING CHANGE: X, Y, Width & Height properties on node class, which were already deprecated, now removed. Use Bounds instead.
BREAKING CHANGE: ValueOfInformationOutput MutualInformation and Reduction properties renamed to support both MI and MEU outputs.
BREAKING CHANGE: Excel files pre 2007 no longer supported by default User Interface connectors.
BREAKING CHANGE: DetectInteralFloats in VariableGenerator now defaults to true
BREAKING CHANGE: DataPartitioning now includes PartitionCount
BREAKING CHANGE: Parameter learning TimeSeries mode now needs to be explicitly set to either Pinned or Rolling with API. Previously defaulted to Pinned.
BREAKING CHANGE: Parameter learning with Decision nodes now requires DecisionPostProcessing to be explicitly set.
BREAKING CHANGE: AutoInsightOptions.KLDivergence changed from bool to enum to allow for future additions.
Version 8.0.0.0 Beta - Sunday, July 8, 2018
.NET API now targets .NET Standard 2.0, with support for .NET 4.6.1+, .NET Core 2+ and more
New JavaScript & TypeScript API
Retracted Analysis for anomaly detection diagnostics
Fix for 'Add nodes from data' with decimal data type (e.g. currency).
Version 7.25.0.0 - Wednesday, September 20, 2017
Fix for MySql .NET connector in User Interface.
Fix in user interface in parameter learning wizard with noisy nodes
Version 7.24.0.0 - Thursday, July 20, 2017
Performance fixes to Impact Analysis and Log-Likelihood analysis with larger numbers of variables.
Version 7.23.0.0 - Wednesday, July 19, 2017
Log-likelihood analysis tool.
In-sample anomaly detection. Helps remove anomalous data from a training set.
API: DataTable default implementation of IEvidenceReaderCommandFactory. Helps create partitioned data sets more easily.
API: kFold cross validation now includes non-callback version.
Version 7.22.0.0 - Saturday, July 15, 2017
Link strengths can now be displayed graphically in the network viewer
Layout in compact mode now takes into account label bounds
Nodes can now be added to groups via selection in the network viewer
Changes made to Compact/Query node views now support undo/redo
Version 7.21.0.0 - Tuesday, July 11, 2017
Node groups. Groups can be colored for display in the User Interface
New compact node view in user interface
New icons in User Interface
Network viewer magnifier
Version 7.20.0.0 - Tuesday, July 4, 2017
New Hierarchical structural learning algorithm. Builds layered groups of variables.
New structural learning wizard.
New parameter learning wizard.
Add nodes from data (VariableGenerator) now reports % of missing data for each column.
Batch query User Interface no longer keeps handle open to data connection after processing.
Structural learning User Interface now supports learning with a subset of nodes.
Auto Insight User Interface now supports filtering test variables.
Search/Score structural learning now supports link constraints.
Version 7.19.0.0 - Saturday, June 10, 2017
Discretized ranges now displayed in UI elements
Bug fix for continuous <-> discrete nodes in Chow-Liu and TAN structural learning algorithms
Link associations: fixed incorrect listing of variables and nodes in User Interface
Version 7.18.0.0 - Tuesday, May 23, 2017
Impact analysis.
Version 7.17.0.0 - Saturday, May 20, 2017
Auto-insight now supports continuous test variables in addition to discrete.
Auto-insight now reports the Kullback-Leibler divergence for discrete & continuous variables.
API now supports KL Divergence for discrete or continuous query distributions.
Version 7.16.0.0 - Thursday, May 18, 2017
Value of information now supports discrete, continuous & mixed/hybrid networks.
New Link strength / association calculator
Value of information for decision graphs
Mutual information calculator now supports mixed discrete-continuous distributions
Entropy calculator now supports mixed discrete-continuous distributions
Kullback-Leibler divergence calculator.
Feature importance (selection) now includes mutual information.
Auto-insight now supports copy to clipboard.
Version 7.15.0.0 - Thursday, May 11, 2017
Chow-Liu algorithm now supports networks with a mixture of discrete & continuous nodes (hybrid)
TAN algorithm now supports networks with a mixture of discrete & continuous nodes (hybrid)
Batch discretization of columns. This improves the performance when discretizing many columns
Fixed IsConstant bug in 'Add nodes from data wizard'. Only affected the user interface.
Added progress and cancellation to 'Add nodes from data wizard' (VariableGenerator in API).
Version 7.14.0.0 - Monday, May 1, 2017
New Search & Score structural learning algorithm
Memory leak in inference engines fixed
Detect integral floats option added to variable generator (Add data from nodes)
New 'Add nodes from data' wizard added to User Interface.
New 'Message Bar' added to 'Network Viewer' in the User Interface that guides new users
Variable generator (Add nodes from data) detects constant variables
Version 7.13.0.0 - Wednesday, April 5, 2017
Mutual information calculator - User Interface
TAN structural learning now supports continuous variables
Chow-Liu structural learning now supports continuous variables
Application shortcuts now uninstalled correctly when previous major versions are installed.
'Add nodes from data' now supports Decision and utility nodes.
Version 7.4.0.0 - Saturday, April 23, 2016
Simple in-memory data store (DataTable, DataRow,...) added to Java API. Allows easier integration of data driven features from R (Data Frames), Python (Data Frames) & Matlab.
Data windowing added to APIs (WindowDataReader, WindowDataReaderCommand).
Version 7.3.0.0 - Monday, April 11, 2016
Default log-likelihood tolerance changed to 1e-6
BREAKING CHANGE: Java IDataPartition interface renamed to DataPartition.
BREAKING CHANGE: Java DataPartition & EvidencePartition moved to com.bayesserver.data.distributed namespace.
Version 7.2.0.0 Alpha - Saturday, March 19, 2016
Fixed csv connection editors in the user interface
Version 7.1.0.0 Alpha - Friday, March 11, 2016
Help added for Decision graphs (LIMIDS)
CSV support added to user interface.
New Oil Wilcatter sample decision graph added.
Walkthrough data now included in an Excel spreadsheet.
Multivariate conditional utility queries (can be used to generate data for profit/cost surfaces).
Decision graph parameter learning.
Utility variables with variances.
New Expert disagreement sample network added.
Decision graphs with both discrete and continuous probability (chance) nodes.
BREAKING CHANGE FROM 6: The default convergence method for parameter learning is now log-likelihood.
BREAKING CHANGE FROM 6: The default prior for parameter learning is now uniform.
BREAKING CHANGE FROM 6: The default initialization method for parameter learning is now clustering.
BREAKING CHANGE FROM 6: ParameterLearningOptions.Tolerance now defaults to null, with the default depending on the Convergence method used (available via ToleranceOrDefault).
Version 6.17.0.0 - Saturday, January 30, 2016
Detect cluster count now supported in the user interface.
Fix for clustered initialization with weighted data during parameter learning.
Version 6.16.0.0 - Friday, December 25, 2015
Fix for disabled menu items in the User Interface, caused by a bug in .NET 4.6.1
Fixed NaN error which occurred during clustered initialization with a variable with zero variance.
Fixed 'Specified argument was out of the range of valid values' error in structural learning caused by zero degrees of freedom in test.
Version 6.15.0.0 - Saturday, October 10, 2015
API: Detect the number of states of a discrete latent variable (cluster/mixture).
API: Convert from a covariance matrix to a correlation matrix.
Version 6.14.0.0 - Sunday, October 4, 2015
Conditional queries such as P(A|B) and P(A,B|C,D) now directly supported in the API and User Interface.
Soft evidence can now be adjusted by existing distributions or queries.
List of errors now available in the Confusion matrix user interface.
Checkbox bindings in 'Add nodes from data' dialog fixed.
Network viewer theme mordernized.
Version 6.13.0.0 - Sunday, August 9, 2015
Performance of multi-variate Gaussians improved.
API: New instantiation methods on CLGaussian, that provide improved performance
Structural learning now correctly supports discrete variables with only a single state
Version 6.12.0.0 - Saturday, June 27, 2015
Avoid 'Covariance matrix is not positive definite' errors during structural learning, with numerical stability enhancements.
Version 6.11.0.0 - Saturday, March 28, 2015
Add multiple states to a variable.
Further improvements to the numerical stability of high dimensional Gaussian distributions.
Gaussian distribution plots can be weighted, to decrease visibility of lower support Gaussians.
Temporal scatter plot - fix for variables plotted on incorrect axes.
Version 6.10.0.0 - Wednesday, March 25, 2015
Query plots now support ordering of variables.
Clustered initialization - performance improvements with large discrete variables.
Clustered initialization - handling of temporal variables improved.
Clustered initialization - improved handling of conditional latent variables.
Version 6.9.0.0 - Saturday, March 21, 2015
Clustering initialization method fixed to work with distributed parameter learning.
Numerical stability of high dimensional Gaussian distributions improved.
Version 6.8.0.0 - Saturday, March 14, 2015
New parameter learning initialization algorithm [Clustering], that often provides better results with fewer EM iterations. By default this is not enabled, to maintain compatibility with version 6.
Rendering of Gaussians (visualizing queries and parameter learning monitoring) improved to handle 1000s of Gaussians.
Initialized state now output at iteration 0 during parameter learning progress notifications.
Random sampling parameter learning initialization fix that provides improved starting covariances.
Version 6.7.0.0 - Tuesday, February 10, 2015
Performance improvements to Relevance Tree inference with large discrete variables.
Fixed incorrect max propagation log-likelihood calculation for certain Relevance Tree inference queries.
Performance improvements to parameter learning.
Fixed occasional incorrect display of the last variable in the Data Map window in the user interface.
Version 6.6.0.0 - Sunday, October 5, 2014
Uniform prior option added to parameter learning user interface
Data tables are now cached by function in the user interface.
New variable argument Node constructor added for creation with variables.
Version 6.5.0.0 - Thursday, September 25, 2014
Automatic check for updates now has limited support for automatic proxy detection.
Fixed Table.getMaxValue in Java API. Previously properties were hidden.
Fixed information columns filter in data map window
Fixed data explorer bug caused when non temporal data was not mapped when temporal data is.
Improved the performance of Node explorer for discrete variables with large numbers of states.
Version 6.4.0.0 - Thursday, September 18, 2014
Improved inference performance with large discrete variables.
Improved parameter learning performance with large discrete variables.
Version 6.3.9141.0 - Wednesday, September 3, 2014
Documentation updates.
Version 6.2.9118.0 Beta - Thursday, August 21, 2014
Added selection likelihood per pane in auto insight viewer.
Documentation updates for new features
Version 6.2.9111.0 Beta - Tuesday, August 19, 2014
Fixed scrolling in auto insight viewer.
Version 6.2.9110.0 Beta - Sunday, August 17, 2014
States are now sorted in evidence explorer.
Performance of auto insight user interface improved.
Minor breaking changes to distributed parameter learning method arguments.
Version 6.1.9099.0 Beta - Monday, August 11, 2014
Uniform discrete priors
Fixed log-likelihood monitoring in distributed parameter learning
Confusion matrix grid view
Version 6.0.9068.0 Alpha - Thursday, August 7, 2014
Save and load Networks from strings.
Fixed MarginalizeLowMemory in MPE mode.
Node, variable and state lookup by name can optionally throw if not found.
Many new methods for easier access to Table and CLGaussian parameters, using Variables and States directly.
New methods to facilitate easier creation of Nodes, Variables and Tables using variable length parameter lists.
Version 6.0.8976.0 Alpha - Tuesday, July 29, 2014
Excel batch queries fix
Batch query create table data types fix
Version 6.0.8931.0 Alpha - Thursday, July 24, 2014
Auto insight decision tree
Version 6.0.8901.0 Alpha - Sunday, July 20, 2014
Fixed check for update code for pre-releases.
Version 6.0.8899.0 Alpha - Saturday, July 19, 2014
Distributed parameter learning
Auto insight
Breaking change to EvidenceReader interfaces.
Version 5.6.8979.0 - Tuesday, July 29, 2014
Fixed check for update code for pre-releases.
Fixed bug in Network.Load
Batch query create table data types fix
Version 5.5.8753.0 - Saturday, March 8, 2014
Fixed query explorer repaint issue.
Zoom slider improved.
Fixed create tables sql generation - state indexes option
Version 5.4.8708.0 Beta - Saturday, February 22, 2014
Fixed bug in the batch query user interface. Discretized variables were not available when constructing a confusion matrix.
Improved the numerical stability of the structural learning algorithm, especially with highly correlated data.
Many new help topics and new feature help completed.
New buttons to facilitate easier adding of default Case, Weight and Time column names.
User interface now defaults to times not being zero based indexes.
New noisy node sample network
New walkthrough - Discovering insight using comparison queries
Version 5.3.8627.0 Beta - Sunday, January 26, 2014
Fixed additional bug in 'Add nodes from data' which lead to an 'Object reference not set to an instance of an object' error.
Improved equal frequency discretization algorithm with handling of integral types.
Improved equal frequency discretization algorithm to more often return the suggested bin count with heavily skewed distributions.
Fixed misleading error reporting when invalid data is detected during concurrent parameter learning.
Version 5.2.8606.0 Beta - Monday, January 20, 2014
Cross validation added to API
Residual plots added to batch query window
Fixed bug in temporal structural learning
Fixed timeout bug when writing large batch query predictions to a data source
Fixed State Not Found Action bug
Fixed bug in 'Add nodes from data' which lead to an 'Object reference not set to an instance of an object' error.
Version 5.1.8329.0 Beta - Wednesday, October 2, 2013
Structural learning
Java API
Version 5.0.8118.0 Alpha - Sunday, September 1, 2013
Alpha expiry date extended
Version 5.0.7738.0 Alpha - Wednesday, May 29, 2013