The keynote today was about Windows 7. I saw a lot of things that I think will improve the average day of most Windows users. I especially liked the MRU list showing up with the icons on the task list. That will save me multiple clicks each day.

I really liked the demos of synchronizing data between different office applications. I work with a lot of remote teams and sometimes Live Meeting doesn't seem to be the best environment.

I also liked the keynote on the Azure services and moving from a local service to one hosted in the cloud. I had seen a similar demo less than a month ago at an NDA event but it was still great to see.

I went to a "lap around Oslo" session and saw some good stuff with the modeling language "M", the tool "Quadrant", and the repository.

Most of this afternoon was spent working and talking to a lot of the other RDs. For me getting to see all these smart people may be the best part of a conference like this.

You can see screencasts and get news from PDC from a variety of sources including microsoftpdc.com so I won't try to capture every detail but instead put down my thoughts on what I learned/saw.

Today has been an Azure day. In the morning keynote Microsoft announced Microsoft Azure, their new operating system for the cloud. I have spent most of the rest of the day in sessions learning more about the cloud and what all it can do now and what it will do in the future. I have some small ideas for the SQL Services storage and how to use it. I just registered for my Azure account so I can test out the CTP.

One of the useful and impressive demos I saw was of a workflow that started out as a local WF workflow and then sent messages into the cloud where the workflow continued and finally messages were sent back to the originating workflow. I can envision scenarios where certain things like payroll or invoicing could be made part of the workflow and companies could reduce costs for commodity items that are not really part of their core business.

In other news the WCF REST starter kit was released on CodePlex. You can access it at http://www.codeplex.com/aspnet/Wiki/View.aspx?title=WCF%20REST. There is also documentation and other information at the WCF Developer Center REST site.

Also LiveID will become an OpenID provider. It is in CTP now but if it works and gains wide spread adoption it will enable a lot more "single sign on" scenarios. This looks like the first concrete step to come out of Microsoft agreeing to support OpenID.

finally, while not PDC related, TJay Belt had an article on duties for on-call DBAs featured at SQLServerCentral.com.

I posted yesterday on Google's new browser named "Chrome". Today I have see a lot of news articles and blog posts about the browser. In general the feeling seems to be that the browser is fast, has a lot of good features, and is squarely aimed at taking market share from Microsoft. Not only will it hurt IE if Chrome gets a lot of play but it will also affect Windows as the browser will become the application platform.

There was a lot of talk about the EULA (you know the thing you never read but click the "I agree" button so you can get on with installing software). In particular the agreement granted Google a license to use anything you used in the browser. There was a wide range of reactions from "this is standard practice to keep Google from being sued, they would never use your stuff" to "that means they can capture and post the details of my on-line banking session to the world. I am deleting this as fast as I can". Whether they were intending to do anything with the data or not is unclear to me but in an article on arstechnica.com Google admitted it was a mistake and promised to change the license to retroactively not grant them any rights.

There is a bigger issue though and that is the Omnibox on each tab. The arstechnica article links to another article on cnet.com that states "Chrome's "Omnibar" can also access all keystrokes a user types, and Google will store some of this information along with IP addresses". This sounds like the kind of scenario I wrote about back in 2005 where AJAX is used to send your keystrokes to a server even if you don't ever actually hit the enter key. I have not found any mention of Google deciding to stop this practice in the future so I guess I will stay away from using Chrome until I have a better feel for what privacy I am giving up. I like the idea of having it suggest information to me, I just don't want Google storing that information.

It seems that yesterday Google leaked information about a new browser named "Chrome" they are building. The blog announcement and comic book with details are interesting. Some of the things I see are also in IE 8 and some seem to be a little different. As both IE 8 and Chrome are in beta it is too soon to start a comparison but here are some of the things I like and dislike about the announcement.

Like:

• Each tab is a separate process so a tab crashing will not bring down the entire browser
• Improved security including a privacy mode where nothing is saved. This will be good for going to web sites that might be legitimate or might be a drive by download site
• Webkit based rendering engine. At least we won't have to be worried about a different way of handling HTML or CSS

Dislike:

• Each tab is a separate process (I know I said I like it above) because processes take more resources than threads. I am not sure I will notice a slowdown switching between processes but that possibility exists.
• New JavaScript engine. I am sure it will be fast and the most standards compliant (whatever that really means) engine out there but it will be different which means one more browser to test on and I am getting lazy in my old age :)

Only time will tell whether this means increased competition for IE, a fragmentation of the open source based browser market, or some combination but I believe that competition will help all the products be better.

Until I see a compelling reason to switch I still plan on doing most of my browsing in IE with FireFox, Safari, and now Chrome installed to help with testing and debugging of pages.

Here is the body of an e-mail that I got announcing that the latest CTP of BizTalk Services has been released and is ready for us to start working with and providing feedback on.

Announcing the BizTalk Services "R12" Release

We're thrilled to announce that the BizTalk Services "R12" Community Technology Preview (CTP) is now available for general use.

"BizTalk Services" is the code-name for a platform-in-the-cloud offering from Microsoft. Currently in active development, BizTalk Services provides Messaging, Workflow, and Identity functionality to enable disparate applications to connect quickly and easily. Combined together in an integrated offering, these capabilities deliver a Service Bus architectural pattern that is immediately usable by applications that need to connect across the Internet.

Many enterprises employ the 'Enterprise Service Bus' pattern to interconnect disparate systems within an organizational domain. Built on Microsoft platform technology, an ESB might include building blocks such as Windows Server, Active Directory, BizTalk Server, as well as the Windows Communication Foundation and Windows Workflow Foundation technologies included in the .NET Framework. "BizTalk Services" extends the concept of an ESB to truly exploit the Internet, for instance by exposing individual service endpoints in a secure fashion or by selectively federating elements of distinct identity systems to facilitate cross-company collaboration.

For ISVs and Solution Providers creating specialized business solutions that enable collaboration and information exchange across increasingly mobile and distributed work-forces, "BizTalk Services" provides the cloud-based platform building blocks to create sophisticated (Internet-) Service Bus solutions with broad reach that could otherwise only be realized by operating dedicated Data Centers of significant complexity - which is often out of reach for both, ISVs and their customers.

Major Changes

With the release of BizTalk Services "R12", developers must update all clients and SDK installations to the new release.

New in R12 - Workflow

The most exciting new capability we've added in the "R12" CTP is Workflow. These new cloud-based Workflow capabilities enable 'service orchestration' from the cloud. This specialized cloud-based, or hosted, Windows Workflow Foundation runtime can orchestrate services that connect to systems in your enterprise, or to systems running anywhere on the Internet via Web services messages. This new power and capability will enable an entirely new set of application scenarios, and we're very excited to see what people will do with it.

In the SDK you will find samples showing how to create and control Workflow instances hosted on the BizTalk Services cloud, including a sample Workflow implementation that monitors the availability of a website and fires multicast events into the service bus indicating the state.

New in R12 - Identity

For R12, the BizTalk Services Identity Service has been expanded and enhanced to enable more flexibility for scenarios demanded by our customers. R12 introduces a new approach for creating, viewing, and managing access control rules. This approach relies on a few key principles outlined below:

* Every Identity Service account owns a Security Token Service (STS).

* An STS is composed of one or more scopes.

* A scope contains zero or more access control rules.

* An STS owner can grant another Identity Service account permission to edit the access control rules in a scope

A practical illustration to clarify:. The Messaging Service owns an STS whose root scope is http://connect.biztalk.net/services/. When you create a new account (newaccount) in the Identity Service, the messaging service creates a new scope http://connect.biztalk.net/services/newaccount. The Messaging Service then grants (newaccount) the permission to create access control rules in that scope. Any communication endpoints hosted there can thus be secured by the owner of the scope. Rules from R11 accounts have been migrated to the "root" scope of the new account.

On the protocols front, we've added several new capabilities for 'REST' services. We now support integration with Windows Live ID and have added RFC2617 Basic and HTTPS/Client Certificate support for acquiring security tokens using simple HTTP GET requests.

New in R12 - Messaging

Connectivity Modes

The most fundamental new feature area in the Messaging service are the new 'connectivity mode' settings on the RelayBinding. Before this release, BizTalk Services clients and listeners always required outbound TCP ports 808 and 818 to be available for connecting to the BizTalk Services cloud for all connection modes except the clients of a listener running with ConnectionMode.RelayedHttp.

In this release we are introducing three different connectivity modes: Tcp, Http, and AutoDetect. The connectivity mode can be set on a static property of the RelayBinding. The Communication\ExploringFeatures\ConnectionModes\Multicast sample shows how. For clarity: 'Connection Mode' defines the type of end-to-end connection that is to be established through the Relay. 'Connectivity Mode' defines how a particular endpoint connects up to the Relay.

The 'Tcp' connectivity mode is the most efficient one and works as in previous releases. The 'Http' mode is new. It creates a volatile FIFO buffer for messages in the BizTalk Services cloud and polls for messages using HTTP 'parked requests'. The Http model exhibits delivery latency characteristics similar to Tcp mode, albeit with slightly higher bandwidth consumption on idle connections. The 'AutoDetect' mode will check whether TCP connectivity is available and will choose 'Tcp' if that's the case and 'Http' otherwise.

The new HTTP-based connectivity option is only effective for the RelayedOneway, RelayedMulticast and RelayedDuplex connection modes. RelayedDuplexSession, HybridDuplexSession, and RelayedHttp (listener only) still require TCP connectivity at this time.

Transport Credentials and Unauthenticated Access

Also, in the "R12" release, the model for specifying the client credentials for the Relay has now been closely aligned with the standard WCF client credentials model. Instead of picking and instantiating token providers, there is now a TransportClientEndpointBehavior that holds all credential information and credential types. The samples in the Communication\ExploringFeatures\RelayAuthentication of the SDK download clarify the use of this new behavior.

We have added a pair of 'WebNoAuth' samples which introduce a new capability that we had a lot of requests for: Unauthenticated client access. When registering a service listener you can now explicitly waive the authentication requirement for clients connecting to your service. This is very useful in Web scenarios where you want to enable any HTTP client to connect to your service and don't want them to authenticate in any way. For the time being we suggest that you always use this new unauthenticated access mode for RelayedHttp services until we release the update for the 'Web' client authentication capability.

For R12, we have omitted the 'Web' (REST) samples for Relay authentication since that area is undergoing some substantial protocol changes. The update for this will be released soon. In the interim, existing applications that were built on a prior release of the BizTalk Services SDK to use the authentication technique shown in the R11 'Web' sample must be modified to use unauthenticated access as shown in the new 'WebNoAuth' sample.

Give it a try

The new BizTalk Services "R12" CTP is online and available now for your use. The SDK is available at http://labs.biztalk.net. If you already have an account for BizTalk Services, your accounts and settings have been migrated to the new environment. If you don't have an account yet, just sign up, download the SDK, and get started creating the new generation of connected applications.

Last night Microsoft announced that the technology preview of Live Mesh is available to a limited number of beta testers. Live Mesh is a way for me to sign up all of my devices (mobile and Mac coming soon) and designate libraries (files) on each one to be synchronized to the others. I also get a secure remote desktop into the machine. This all runs in the cloud and uses your Live ID and SSL to secure the data. I can see this enhancing/replacing the array of tools I use now for synchronizing and remote access to my machines and data. I am excited to start working with the technology preview and see how this will work when I am on the road next for remote access to my home PC. Right now if I need to get to it I can have someone log into it and I use the remote help option as an easy way to do remote access without having to configure firewalls and worry about having a connection from the Internet to that system. The downside of that is that someone has to request the help and then allow me to have control so it takes someone being available. From what little I have seen about Live Mesh I don't have to have someone at the device to connect to it securely so it will make it more convenient.

Of course at the top of my wish list is to have Live Mesh work with CardSpace but since I can set up my Live ID to use CardSpace I don't know that it is that big of an issue.

When the SDK comes out later it will be interesting to see how Microsoft delivers on their promise of having the same API on the client and in the cloud. If the API looks just like the current .NET Framework that would go a long way to making this the "killer application" of clould computing even though it is really a platform to create killer apps. If all of the .NET programmers can just write their applications and then at some time decide to move them from an on premise app to a being hosted in the cloud that would make this platform extremely useful and really enable it to take off. If I have to write my applications in a new way with different constructs and they only run in the "mesh" then I don't know if I would start targeting this platform for general user applications. There are a class of applications that use collaboration heavily and could use the mesh immediately but I don't know that writing your typical data entry and reporting line of business application would gain enough benefit from the mesh to justify a new programming model and the drop in productivity that it would entail.

For now I will just try to explore what I can do with the technology preview and post my experiences.

You can find more information about Live Mesh at the web site http://www.mesh.com. There are also the following resources available:

- Watch the interview with Ray Ozzie introducing Live Mesh on Channel 9 (link: http://channel9.msdn.com/ShowPost.aspx?PostID=399578) 

- Watch an interview with Abolade Gbadegesin on Live Mesh Architecture (link: http://channel9.msdn.com/ShowPost.aspx?PostID=399577) on Channel 9
- Watch the demo of the Live Mesh application on Channel 10 

Some good background on Mesh can be found at http://blogs.msdn.com/livemesh/