1 Comment


Yesterday evening I found that McDonald's is discriminating against the car-impaired. I am currently in Seattle for the Microsoft TechReady conference. After a photo walk with two colleagues we were getting thirsty and hungry. I managed to stay out of McDonald's for the almost four weeks that I have been in the States, but the McDonald's two blocks away from the Space Needle was a bit hard to resist.

We arrived there at 23:02 and the "restaurant" had just closed. However, the drive thru (yes, that seems to be the correct spelling here) is open 24 hours per day. But what do you do when you don't have a car? You walk through the drive thru. We weren't carrying enough metal to trip the detector in front of the ordering pole. So we couldn't state our order. We preceded to the pay window to order there. The guy there looked very amazed and said he could not and would not serve us. WTF! "Why not?" we asked. "Because you don't have a car", he replied. Disappointed we left. Next time, I will try riding a bike through the drive tru.

In the mean time I will boycot McDonald's for a while.

PS: This is just for my own good as I have the feeling that I gained a couple of kilos here in the US.

It depends on the situation if I self-identify as a geek or not. Today, I thought it would be fine, so I signed up for the Geek Dinner organized by Scott Hanselman in Bellevue, WA.

Quite a lot of people showed up. I went there with my colleague Erwin van der Valk. He was a Development Consultant, like I currently am, at Microsoft Services in the Netherlands. Erwin now works on the Patterns & Practices team at Microsoft in Redmond.

I took this picture of the entire group:

Microsoft Geek Dinner 

After dinner, a large portion of the group went to the Rock Bottom Restaurant and Brewery in Bellevue to continue the conversation. I had some really interesting discussions, over beer, with guys from several different product teams and a fellow MCS consultant based in Denver, CO. And not even all about Microsoft technology 😉

Too bad, I won't be able to attend this event in the near future again, unless I just happen to be in the neighborhood.

You probably already noticed that Microsoft has enabled access to the source code of parts of the .NET Framework 3.5. Here is the announcement of availability in case you missed it. And there are detailed instructions on how to enable this as well.

This morning I started stepping into the Windows Presentation Foundation. Before I knew it I was deep down into the internals of dependency properties and the Dispatcher object. The dispatcher has a nice old-fashioned message pump. I stepped into the dispatch of a message call onto a non-UI thread. Even though I already knew this, it is still startling to see that this translates into old-fashioned User32 Windows messages being send.

What really surprised me was to see a bit of VB code in a file that is called ExceptionFilterHelper.vb. According to the comment this file was needed because C#, unlike VB, does not support exception filters. However, the filter always returns true. According to the comment in the source code it is a "dummy stub to keep compiler happy, will not be replaced".

Something weird is going on here. The debugger shows I am in WindowsBase.dll!System.Windows.Threading.ExceptionWrapper.TryCatchWhen. The source code file shows the class MS.Internal.Threading.ExceptionWrapper.TryCatchWhen. Reflector can only find the former class. The disassembly of this class shows source code that is very different than what the MS source code server gives me.

This doesn't make much sense to me. Is the source code for this class deliberately hidden or is there some error in the source code lookup?

Since I haven't had my lawyer take a look at the license agreement that governs the source code access, I will not add a screen shot of the VB code 😉

1 Comment

Best wishes to everyone for 2008!

In the first day of the new year, I've released version of my Flickr Meta Synchr tool. You can always find the latest version on CodePlex.

I had been working on this new version for a while now, but didn't get around to finishing it. Improvements in this new version:

  • Added much better activity logging. The activity log can now be shown in an additional window and is persisted to disk.
  • Added the option to match pictures on title and filename. This is useful when images have been timeshifted and cannot be matched on date taken.
  • Bug fixes. Improved stability when corrupt image files are encountered. Fixed GPS roundtripping bug.
  • Should run better on 64-bit versions of Windows XP and Vista.
  • Solution is now built in Visual Studio 2008 without the need for any additional WPF extensions.

Here is a screenshot of version

Flickr Metadata Synchr v0.9.0.0

During the lunch break Carlo and I walked around the Microsoft Corporate Campus in Redmond. It is autumn here as well. Lots of deciduous trees are shedding their leaves. The campus now shows lots of shades of red and yellow on leaves.

I took some pictures during the walk. You can see some below. Click on the images to see larger versions on Flickr:

Autumn in Redmond 

Carlo on the Microsoft Way Carlo standing at a Microsoft Way street sign.

Building 42 Building 42. Home to the .NET Framework teams.

Building 42 The other side of building 42.

Microsoft Conference Center The Microsoft Conference Center.

More pictures can be found on my Flickr page.

1 Comment

Microsoft used the keynote in the SOA & Business Process Conference in Redmond to present its vision on the future of Service Oriented Architecture on the Microsoft platform. That vision and the wave of technology that will come with it, is codenamed "Oslo".

MVP Charles Young has a solid write up in a blog post called "Microsoft 'Oslo' - the vNext SOA platform". No need to repeat all that here.

Long time Microsoft watcher Mary Jo Foley is very critical in her post called "Microsoft talks SOA futures but not dates". Mary Jo ends with "Microsoft has been struggling to prove to the market that it has a real SOA strategy. While the Redmondians are talking the right talk, the company is still a ways away from walking the SOA walk. Will customers wait or run off with other SOA vendors before Microsoft rolls out more than just a piecemeal SOA strategy?"

"Oslo" is obviously a Grand Vision. It will take a couple of years before this next wave of Microsoft technologies will ship. I thought that after the Longhorn reset/WinFX debacle and the "Whidbey" delays, Microsoft would not attempt to align so many technologies again in the future. But it is! "Oslo" comprises of at least:

  • BizTalk Server "6"
  • Visual Studio "10"
  • .NET Framework "4"
  • Systems Center "5"
  • BizTalk Services "1"

Some of the stuff presented reminded me of the grand WinFX, especially WinFS, vision that Microsoft presented at PDC03. We all know that WinFS never RTM-ed, despite enormous effort (many, many man years) put into it by Microsoft. Especially the term "Universal Editor" for the "Oslo" integrated modeling tool gave me the creeps. Sounds too much like: One tool to rules them all. One tool that spans the entire application development lifecycle: from its inception to its deployment.

Here are some screenshots from the new "Universal Editor" modeling tool that was demoed during the keynote:

Microsoft Oslo Universal Editor

Microsoft Oslo Server List

Microsoft Oslo Application Verifier

1 Comment

One of my post popular blog posts deals with problems I was having with the stability of Windows Vista. I had major problems hibernation and sleep modes on my Dell PC. My PC never reliably went to sleep or hibernation and crashed either during going to sleep or when waking up. A lot of other people encountered similar problems. This blog post has 66 comments so far.

No driver update or Windows update solved my problem until October 9, 2007.

Around that time Microsoft released a set of reliability updates that are going to be part of SP1. There are supposedly also going to show up in Windows Update, but I manually downloaded and installed them ahead of time. Note that I am not running the beta of SP1.

If you encounter similar problems with Windows Vista, I highly recommend that you install these updates as well (if you haven't got them through auto update yet):


This update resolves some compatibility and reliability issues in Windows Vista. By applying this update, you can achieve better reliability and hardware compatibility in various scenarios.


This update resolves some reliability issues in the USB core components on the Windows Vista operating system.

1 Comment

Scott Guthrie did a major announcement on his blog yesterday: Microsoft will be releasing the source code for most .NET Framework libraries with the release of Visual Studio 2008. There will even be integrated support for debugging into framework classes and on-demand dynamic downloading of source files and debug symbols in Visual Studio 2008.

This is great news for .NET developers and a major step forward for Microsoft in my opinion. In and by itself it is enough reason to warrant an upgrade to Visual Studio 2008. In fact, I can think of no reason to keep using Visual Studio 2005 after the release of VS2008.

The source will be released under the Microsoft Reference License which basically means you can view and debug but not change or reuse the source code.

If you want a more liberal license you can look into Rotor aka the Shared Source CLI. Rotor was Microsoft's first effort for open sourcing a .NET CLI implementation. But Microsoft does not guarantee that Rotor has exactly the same codebase as the real .NET Framework.

Check out the full details and screenshots of VS2008 integration on Scott's blog.

1 Comment

I had taken this online typing test before on a laptop while sitting on the couch. Tonight I was sitting behind behind my desktop computer and decided to retry the test, because Dennis posted his results.

It feels a bit like a "who's got the biggest" contest, but here are the results of my first attempt:

Number of words typed: 219
Test duration: 3 min
Speed: 73.2 words/min. (366 keystrokes/min.)
Error penalty: 12
Accuracy: 94.5%

Not entirely happy with this, I retried the test. Every time you take the test a different text appears, so retrying is not cheating ;). These are the results of my second try:

Number of words typed: 245
Test duration: 3 min
Speed: 81.6 words/min. (408 keystrokes/min.)
Error penalty: 19
Accuracy: 92.2%

Beat this, Dennis! (Shouldn't be too hard 🙂