diff --git a/docs/start/release-notes.md b/docs/start/release-notes.md
index eae54e1..b663d4d 100644
--- a/docs/start/release-notes.md
+++ b/docs/start/release-notes.md
@@ -4,70 +4,111 @@
 toc: ../toc.json
 ---
 
-## Brooklyn Version 0.5.0 RC1 (0.5.0-rc.1)
-
-API Tidy on top of M2, using `init()` as the method to override when building composed Application and Entity classes.
-
-There are known issues with Whirr clusters and the Cloud Foundry example (moving) which will be resolved for RC2.   
 
 
-## Brooklyn Version 0.5.0 Milestone Two (0.5.0-M2)
+## Brooklyn Version 0.6.0 
+
+* Introduction
+* New Features
+* Backwards Compatibility
+* Community Activity
 
 ### Introduction
 
-This milestone release includes many big features, and brings us much closer to a 0.5.0 release.
+This release includes many big features. It incorporates a lot of improvements and feedback from our community. Thank you!
 
-It incorporates a lot of improvements and feedback from our community. Thank you!
+Thanks also go to Brooklyn's commercial users. Brooklyn continues to be used in many exciting projects, including managing distributed database clusters, and cloud-enabling financial service applications (not to mention the custom PaaS offerings, big data clusters, and three-tier web-apps that Brooklyn was being used in at Version 0.5.0).
 
-Thanks also go to Brooklyn's commercial users. Already Brooklyn has been adopted into some very exciting projects including controlling custom PaaS offerings, big data clusters, and three-tier web-apps.
-
-Work on this release can be tracked using Brooklyn's GitHub issue tracker:
- 
-* [github.com/brooklyncentral/brooklyn/issues?milestone=1](https://github.com/brooklyncentral/brooklyn/issues?milestone=1)
-
-And via the mailing lists:
+For more information, please checkout [brooklyncentral.github.io](http://brooklyncentral.github.io), and the mailing lists:
  
 * [brooklyn-dev@googlegroups.com](http://groups.google.com/group/brooklyn-dev)
 * [brooklyn-users@googlegroups.com](http://groups.google.com/group/brooklyn-users)
  
 ### New Features
 
-The major changes between M1 and M2 are:
-
-1. Entities have been separated into an interface and implementation, rather than just a single class. Construction of entities is now done using an EntitySpec, rather than directly calling the constructor. This improvement is required to simplify remoting in a distributed brooklyn management plane.
-
-2. Downloading of entity installers is greatly improved:
-	* More configurable, with ability to specify URLs in the brooklyn configuration files or to override in code.
-	* Will fallback to a repository maintained by Cloudsoft, so if an artifact is removed from the official public site then it will not break the entity. 
-	
-		See [downloads.cloudsoftcorp.com/brooklyn/repository/](http://downloads.cloudsoftcorp.com/brooklyn/repository/)
-
-3. Support for running applications across private subnets.
-
-4. Policies can now be re-configured on-the-fly through the REST api and through the web-console.
-
-5. Some entities now support configuration files being supplied in FreeMarker template format. These include JBoss AS7, ActiveMQ and MySql. More will be converted to use this pattern.
-
-6. Several new entities have been added, including:
-	* MongoDB
-	* Cassandra
-	* RubyRep
-	* DynamicWebAppFabric
+*The changes listed here compare Brooklyn 0.5.0 and 0.6.0, and incorporate changes from the 0.6.0 milestone releases. Items marked **&sup1;** are new since 0.6.0-M2*
 
 
-### Backwards Compatibility
+1. Locations are now constructed using a `LocationSpec`, rather than calling the constructor directly. This improvement is required to allow the Brooklyn management plane to track locations and persist their state.
 
-For upgrading from 0.5.0-M1 to M2:
+2. Location specific and advanced features can be implemented using location extension points. The extension mechanism allows client-code to call `location.hasExtension(extensionClass)` and `location.getExtension(extensionClass)`.
 
-1. Entity classes have been renamed, e.g. MySqlNode is now an interface and the implementationis MySqlNodeImpl.
-	* The minimum change for this to work is to update your references to include the Impl suffixes. However, that will result in deprecation warnings.
-	* The recommended approach is to use the EntitySpec when constructing entities. A good place to start is to look at the updated example applications.
+3. Hazelcast datagrid implementation added. (NB: this is still considered experimental.)
 
-2. The default username for provisioning with jclouds has changed to use the name of the user executing the Brooklyn process. 
-	* Java's `System.getProperty("user.name")` is used instead of 'root' or 'ubuntu'
-	* Usernames can be overridden in `brooklyn.properties` or using system properties.
-		For example, by entering '`brooklyn.location.named.acmecloud.user=root`' in `brooklyn.properties` or using the command syntax `-Dbrooklyn.location.named.acmecloud.user=root`.
-		
-		'`brooklyn.jclouds.aws-ec2.user=root`' could also be used to apply `user=root` to all aws-ec2 VMs. 
+4. A datagrid can be used to store the entity/location state within Brooklyn. This feature is currently disabled by default (i.e. storing to an in-memory pseduo-datagrid); work will continue on this in subsequent releases.
 
-3. Some deprecated code has been deleted. All of this code was commented in 0.4.0 with text such as "will be deleted in 0.5".
\ No newline at end of file
+5. Brooklyn now tracks location and application usage, and records it in the datagrid. More generally:
+	* Create/destroy of VMs (through location manage/unmanage) is tracked
+	* Application start/stop/destroy is tracked
+	* Current and historic usage information can be retrieved through the REST API.
+
+6. Improved Support for Chef, significantly improving support for using Windows machines. 
+
+7. Several additional clouds are supported, including:
+	* Abiquo (see also [this blog post](http://www.cloudsoftcorp.com/news/as-simple-as-abc-the-abiquo-brooklyn-catalog/))
+	* Google Compute Engine
+	* *See also:* Community Activity, below.
+
+8. Several new entities have been added, including:
+	* BindDnsServer
+	* Jetty6Server
+	* MongoDB replica sets
+	* BrooklynNode (for Brooklyn bootstrap deploying Brooklyn)
+	* New PostgreSQL entity (via Chef)
+
+9. A new [location-metadata.properties](https://github.com/brooklyncentral/brooklyn/blob/master/locations/jclouds/src/main/resources/brooklyn/location-metadata.properties) file allows metadata values (e.g. ISO-3166, lat/lon coordinates, etc) to be added to and overridden. Metadata in `brooklyn.properties` overrides metadata in `location-metadata.properties`.
+
+10. GUI Improvements: Better ability to drill down on tasks, auto-updating entity-tree with status indicators, faster status updates, and better performance over slow links.
+
+11. Improved entity and machine restart logic&sup1;, to handle failed components and dead-on-arrival&sup1; machines even better.
+
+12. Improved onboarding, with the creation of a Maven Archetype for a sample application, (which can be used as a template for new applications) and updates to the [Getting Started brooklyn.properties](/use/guide/quickstart/brooklyn.properties) to get HP Cloud and SoftLayer users onboard faster.
+
+13. Sandbox: Support for rolling out complex apps described using CAMP YAML. This simplifies writing application blueprints, as YAML config files are more accessible and easier to write than Java code. &sup1;
+
+14. Big improvements to the Cassandra integration, including new tutorials for high-availability and wide-area (inter-cloud-provider) cassandra clusters.&sup1;
+
+15. The policies for an entity can now be defined in in an entity's spec. (Simplifying adding policies to entities, and requiring less code to be written.) &sup1;
+
+16. Added support for a cluster to spread its members across a provider's availability zones (e.g. in [AWS](http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-regions-availability-zones.html)), to give high availability for that cluster. &sup1;
+
+17. Improved the Brooklyn load balancer abstraction, so that it can be used with things like the CloudStack load balancer or AWS elastic load balancer. &sup1;
+
+
+### Backwards Compatibility  
+For upgrading from 0.5.0 to 0.6.0, existing entities/applications/policies will still work, provided they did not previously have deprecation warnings.
+
+Some additional code has been deprecated - users are strongly encouraged to update these:
+
+1. Use `LocationSpec` (and `managementContext.getLocationManager().createLocation(LocationSpec.spec(MyLocClazz.class)`), rather than calling location constructors directly.
+
+1. In brooklyn.properties Locations must now be specified as `brooklyn.location.*`  (e.g. `brooklyn.location.jclouds.*` rather than brooklyn.jclouds.*). Property names should use CamelCase (e.g. `brooklyn.location.jclouds.publicKeyFile` instead of brooklyn.jclouds.public-key-file). Using the old conventions will give deprecation warnings in the logs.
+
+1. Use `@Effector` and `@EffectorParam` for annotating effector methods in an entity (rather than `@Description`, `@NamedParameter` and `@DefaultValue`).
+`location.getName()` has been renamed to `location.getDisplayName()`, to be consistent with Entity. The `location.getChildLocations()` and `location.getParentLocation()` have also been renamed to `getChildren()` and `getParent()` respectively.
+
+1. If overriding a config key in an entity to change the default value specified in a super-type entity or constant, then use `ConfigKeys.newConfigKeyWithDefault(...)`.
+
+1. Use `ConfigKeys.newStringConfigKey(...)` and similar methods, rather than using `new ConfigKey<String>(String.class, ...)`
+1. `ListConfigKey` has been deprecated because it no longer guarantees order; instead use `SetConfigKey`. This is a consequence of efficiently using a datagrid to store the data.
+
+1. Use the new `FeedConfig.onFailure()`, `FeedConfig.onError()` and `FeedConfig.checkSuccess()`. For example, This is now called to determine if a http 404 is a success or failure (default: failure), and if an ssh non-zero exit code is a failure (default: yes).
+
+1. Some package/class changes, such as:
+	* `brooklyn.util.NetworkUtils` renamed to `brooklyn.util.net.Networking`  
+	* `brooklyn.entity.basic.lifecycle.CommonCommands` renamed to `brooklyn.util.ssh.CommonCommands`  
+	* `brooklyn.util.MutableMap` renamed to `brooklyn.util.collections.MutableMap` 
+
+1. Logging has been updated. Please see [logging.md](/docs/dev/tips/logging.html).  Existing custom logback.xml configuration may no longer work (e.g. brooklyn logback-defaults.xml no longer exists).
+1. `EntitySpecs.spec(...)` is now deprecated; use `EntitySpec.create(...)` instead.
+1. `JcloudsLocationCustomizer`: method signatures have changed. For those extending `BasicJcloudsLocationCustomizer` as recommended, existing code will work but should be changed to use the new signatures.
+1. Deprecated `new BasicConfigKey(key, defaultValue)`. Instead use `ConfigKeys.newConfigKeyWithDefault(key, defaultValue)`.
+
+There have been some other minor changes, where code is now deprecated - all such changes are clearly marked in the javadoc, indicating the recommended approach.
+
+### Community Activity
+
+[Cloudsoft](http://www.cloudsoftcorp.com) has created Locations for IBM SmartCloud Enterprise and Ravello (repos: [github:cloudsoft/ibm-smartcloud](https://github.com/cloudsoft/ibm-smartcloud), [github:cloudsoft/brooklyn-ravello](https://github.com/cloudsoft/brooklyn-ravello)). This code is public, subject to the Cloudsoft developer licence.
+
+### 0.6.0-M2 - GA Changes
+Features marked **&sup1;** (Features 11, 13-17, above) were added in the last push to 0.6.0 GA.
\ No newline at end of file