October 2009 - Posts
The last couple of days I’ve been the guest of Microsoft UK at their Thames Valley Park HQ in Reading. About a hundred MVPs from across the UK and Ireland came together to learn from each other, find out about upcoming technology, to network and socialise. Our schedule began with what I had mentally dismissed as a Microsoft propaganda exercise, but in fact turned out to be perhaps the most significant session of the entire programme. Karina Gibson spoke about how Microsoft is supporting the education of parents into online safety by training employees to deliver an internet safety presentation and provide useful links and resources to parents at local schools. The presentation talks about the benefits and safety concerns of the internet as well as introducing basic information on how technology such as social networking and instant messaging is used by young people. It is designed to build confidence for parents in understanding the technology as well as a providing guidance and practical advice to safeguard their children and where to go for help.
A thought provoking session that spilled over into the after dinner conversation. A number of the MVPs present had personal stories to tell, which were clearly not easy for them to share, but an incredible willingness to share is what distinguishes MVPs and is why Microsoft seeks them out and rewards them. The value of these discussions, made all the more real because they are personal, is incalculable. The people concerned know who they are and I’d like to thank each of you for your openness, it has massively increased my personal understanding of the problems our youngsters face on the internet from threats like cyber-bullying and grooming. This highlights the value of community and is reason enough to attend any community event. The “geek stuff” is why we all go, but the real value is derived from the human interaction.
Today, for the first time in years, the machinery at Lancaster University suddenly ground into action and sent me a text message informing me of a “Possible Sudden Storm Commencement”.
[Update 18:15 - Lancaster University emailed to say it was a false alarm caused by "man made disturbances at the field site"]
This is not just any storm, this is Space Weather.
In rough terms, our sun has an 11 year cycle of activity. At the peak of the cycle, we see lots of sun spots and occasionally, in what is known as a coronal mass ejection, the sun spits out a puff of plasma consisting mainly of subatomic particles. Sometimes, these CMEs head straight for the earth at an average speed of about 500 kilometers per second and when they arrive, they interact with the Earth’s magnetic field. The result is that some particles get trapped in the Earth’s magnetic field and they travel down the towards the poles, where the field lines are close together and almost vertical. When the particles reach the ionosphere, they interact with (ionise) gasses and cause the gas to glow, resulting in one of nature’s most spectacular sights that we know as the Aurora Borealis 
(northern lights) or Aurora Australis (southern lights). This is also known as a magnetic storm and the magnetic activity can interfere with communications, satellites and electrical equipment here on Earth.
Recently, the sun has been quiet. Too quiet. The last solar maximum was back in 2001 and the expected increase in activity hasn’t happened – yet. Usually in nature, when things go ‘out of whack’ there is a compensation and some scientists are predicting that the next maximum, expected in 2012, will be the largest since 1859 when the Northern Lights were visible as far south as Rome.
So today’s AuroraWatch Red Alert is significant, it may be the re-awakening of the sleeping giant. I have seen the Auroa Borealis twice in my life, once in 2001 in Yellowstone National Park, Wyoming USA and the second time in 2002 from the top of Eglwysilan, South Wales. I’m really hoping to make it a hat trick.
If you want to see the Aurora, you’ll need to be determined. If you are anywhere near street lamps or city lights, forget it. You’ll need to get to a remote location away from light pollution so you have dark skies and you can see plenty of stars. It’ll also need to be a clear night and of course you’ll have checked AuroraWatch to see if an display is even likely. Then, if you’re lucky, you might just see some activity in the northern sky. When I saw the Aurora from Wales, I was on Eglwysilan mountain above Pontypridd and the display was faint milky-white rays with the occasional tinge or orange.
Those cunning Japanese have figured out a great way to commemorate Windows 7 General Availability: The Windows 7 Whopper!
I’ve been having all manner of problems lately building my Visual Studio projects, and in particular, building them on my TeamCity build server. Builds that worked OK in Visual Studio would fail in my TeamCity build agent, even though the build agent was running on the same machine as the same user. Things that have worked for years suddenly freaked out, including:
- Compiling satellite resource assemblies (can’t find AL.exe)
- Compiling DocProject projects (weird threading issues).
- Building and running MSTest projects.
It seems like something odd is going on in the world of MSBuild. I don’t really understand why this has been necessary. My builds have always worked in the past – why not now? I never needed this SDK before, just having Visual Studio installed was enough. So why now do I need to install this SDK? It seems to me like something has updated MSBuild ‘on the quiet’ and broken it. To be sure, some of the MSBuild scripts are in different locations now to what they used to be, why did MSBuild start looking in the new locations? What caused that?
These problems were solved by downloading and installing the incongruously-named Windows SDK for Windows Server 2008 and .NET Framework 3.5. Even though I’m running on Windows 7.
There are some ‘gotchas’ when installing this update on top of Visual Studio, so be careful. You need to re-apply Visual Studio Service Pack 1 afterwards – but it does solve the build problems.
I learned something about .NET assemblies, COM Interop and strong-naming today that I thought would be worth documenting here for later when I forget ;-)
For purposes of COM Interop, the location of a .NET assembly and whether or not it is strong-named affects the visibility of the COM interfaces. We have three distinctly different situations:
- Assembly is strong-named and installed in GAC
- Assembly is strong-named but not installed in GAC.
- Assembly is not strong named
These conditions determine the visibility of the assembly to COM components. When a COM component tries to load the assembly (via COM Interop) then Fusion applies the normal .NET rules for assembly binding, which results in one of the following outcomes:
- Assembly is in the GAC. The COM interface is globally available to all COM components. Fusion can always find assemblies in the GAC and they take priority over local copies. By definition, the assembly must be strong named to be in the GAC.
- Assembly is strong-named but not installed in GAC.
- If the assembly was registered with the /codebase option, then the COM interface is globally available.
- Otherwise, the COM interface is only available if the assembly is in the same folder (or a subfolder) of the COM component trying to load it.
- Assembly is not strong named. When there’s no strong name, /codebase can’t be used. The COM interface is only available if the assembly is in the same folder (or a subfolder) of the COM component trying to load it.