Michael Howard is on a security team at Microsoft. He is the author of Writing Secure Code, and he spoke to us about the security scene at Microsoft.
The #1 agenda at MSFT used to be “app compat” (NOTE: There seems to be a sub-culture at MSFT that uses different terms :), but now security has taken over the top spot.
To enforce this, Michael and his team came up with a new software development process. He detailed this in front of us all.
I asked him about AOP, and he said that he hasn’t thought about it.
To me, enforcing security policies is a perfect example of AOP. A lot of the AOP work that you can get into places, starts off with this style. You write some nice aspects which modularize your policy, and then developers aren’t able to trip up.
This should be the same way in MSFT. If I develop something which doesn’t meant the secure code guidelines, when I do a build, I should get a warning.
This is a much better idea than going through code at the end of the process.
Of course, these policies may not be able to 100% capture the cross-cutting security concern, but they can surely do some of it!
Come on Michael, take it to the next level ;)
