I can relate to this article. Paul is really down to earth. I especially like his comments on cube vs. office. I am definitely one of those people that needs quiet space and find it extremely upsetting when I get interrupted at the office or at home.
Check it out…
http://channel9.msdn.com/ShowPost.aspx?PostID=13369
The Xbox 2 specs speak for themselves.
The compiler technology hints at something interesting. It appears as though Microsoft is going to have some really interesting open, intermediate code that will allow the code generation to be really flexible. We need compilers like this if we are going to move into the future.
Very exciting times…
There is a huge problem with email lists. People don’t know when they are sending to one. Programs that work with address books don’t know how to recognize them. The end result is that people make mistakes and things get posted to lists (unsubscribes, vacation messages, etc.)
This is a great example of a ‘design hole’. SMTP/RFC822 email is so flexible, it doesn’t differentiate any of the delivery endpoints for email. Unfortunately, this makes it impossible for programs to know that an email address belongs to a list.
IMHO, if I had the chance to do it all over, here is what I would do.
I would make email addresses for lists or any address that the user should be careful of sending to something like this: sales%somecompany.com. That way the user’s email program and any of the programs that used the contacts could easily avoid those addresses or treat them differently.
And I got to wondering.. is there a C# compiler that compiles down into x86 or some target machine language? Some of you may go “why”? Well, for the same reasons that GCJ exists. Sometimes you don’t want to rely on Java Ver X.Y to be around. Sometimes you just don’t want to take the bytecode performance hit.
Is there a GCJ equivalent for C#?
After looking at the DTrace stuff I saw earlier, Solaris 10 actually looks like an OS I would run!
http://wwws.sun.com/software/solaris/10/
There are a lot of interesting new technologies in there. Lets hope they really go back to their roots and create a developer CD that allows people to do C/C++ using their tool chains. When Sun first shipped, I believe the compiler came with the OS.
You know, with this, and a few other Real-Neat-Things out there, I haven’t been this excited with computers in quite a while.
This paper describes an interesting technology that Sun had been working on for quite some time and is now in production. The technology allows an IT admin to do deep performance analysis on parts of the system without having to instrument the system before hand (or unbundling it afterwards). What is also really really interesting, is that the system can be built for this without impacting performance. They call this the “zero probe effect” and it can be applied to the running kernel or the user-land applications. Very cool! Read the paper for the details, but essentially it is clever binary instrumentation.
http://www.sun.com/bigadmin/content/dtrace/dtrace_usenix.pdf
It’s interesting to see Sun finally attempt to compete with other kernels given their slow reaction over the past 6 years. This isn’t something you can just slap into a kernel.
Another interesting thought is what the implications of this are. When you have tools to work with finished code, neat things can be done.
