John Dvorak has written an interesting column on Rajendrasinh B. Makwana. It's a little bit Xenophobic, a little bit protectionist and it also makes some good points that I've often wondered about.

When this story broke, I wondered why FNMA would hire H1-B contractors. I can understand HP or Intel to some extent. But FNMA is not doing anything that average U.S. programmers cannot do. What special skills did Makwana have? Perhaps his best skill was his pay rate?

I guess our own government is willing to "go foreign" in order to avoid U.S. wages? That certainly gives me warm and fuzzy feelings. Particularly since my tax dollars and your tax dollars are being delivered to this outfit by the trainload full.

Does make one think...

The answer is a resounding NO!

1. I have found an inconsistency between BlogEngine.Net and ScribeFire. I just happened to see the problem today.
2. If I post an article using ScribeFire, everything appears to be correct on the blog.
3. Next, if I decide to edit the entry in BlogEngine.Net's editor, I get two copies of the article.
4. If I delete one of the articles, both copies get deleted, however, you cannot see this from the browser where you are editing.
5. If I visit the blog using a different browser, I cannot see either article.

I chalked this up to some operation I was performing incorrectly. However, I think it's a problem in the BlogEngine.Net code and how it implements the blogging protocol BlogML.

I've discovered the source of a performance frustration I've had with VS 2005 ever since I started using it at work. I didn't realize there was an option to turn off background VSS tasks. Occasionally, our servers at work really slow down. When this happens, VS 2005 seems to lock up inexplicably. It turns out VSS tasks are running often in the back ground. Really often. About every five to 15 seconds it seems, based on IP logging. I have a solution with about 80 components. The overhead to continuously update the VSS icons is considerable. The figure belows illustrates what to change:

 

I hope this little discovery may help you as well.

I like programming with WTL and ATL. It's straight forward, clean and quite frankly, simple. I've worked a little with attributed ATL projects and attributes make certain functions a lot simpler.

I needed to recode an ActiveX component recently to work around Vista's IE protected mode. I decided to work with attribtued ATL. Since several dialogs were involved, I decided to drag in WTL to simplify the dialog code. That's when the problems began.

I didn't know this, but attributed ATL projects drag in atlstr.h, via the header file atlplus.h. You need to define _WTL_NO_CSTRING before including atlmisc.h. Otherwise, your component will attempt to compile WTL::CString and ATL::CString. The results are "unpredictable", but generally not fun to resolve.

Here's a reference: WTL CString Class Implemented with the Standard C++ Library

I've reproduced the key items:

1. WTL::CString is not compiled if you #define _WTL_NO_CSTRING before #include <atlmisc.h>
2. You will get ATL::CString support if you #include <atlstr.h> before #include <atlapp.h>
3. When ATL::CString is supported, WTL::CString should not be compiled.
4. If rule 2 does not apply, you will get WTL::CString support for classes defined after #define _WTL_USE_CSTRING
5. If rule 2 does not apply, you will get WTL::CString support for classes defined after #include <atlmisc.h>

 

Here's the order you need to use. Other headers should follow these four.

Groklaw does us all a stupendous service. I find Pam Jones' writing fascinating.

This article is about a large company, attempting to defend its intellectual property, versus a boy who is being sued for $1 Million for stealing seven songs.

There are some substantial constitutional issues in this case. You should take twenty minutes and digest what is going on with this case. Sony BMG Music v. Tenenbaum