It has been a crazy couple weeks, but I think things are now starting to get into order. As I mentioned in a previous entry, I moved baus.net last weekend. The transition wasn't as smooth as I hoped, because I had to rush to get up to the colo to retrieve my server. I could have used another week to prep, but the added pressure of trying to get my site back up and running with minimal downtime actually made the transition happen.
I was also contacted by a C++ student who is interested in getting the SwitchFlow framework ready for Open Source, and I decided to take him up on his offer. I'll let him introduce himself if he wants. I'm doing this for a couple reasons:
- I want the code to be something I can feel comfortable with.
- I want fellow software designers to actually use the code.
- I don't want to spend a lot of time answering questions about how to download and build the code.
I also upgraded to the latest version of Libevent. While it built without a hitch, of course it seems to have caused problems running the proxy server. The event distribution layer in SwitchFlow could use some work, but I was hoping I could put that off.
I sent a draft of an article describing my C++ pooled_list to the ACCU Overload journal. I will admit I have had poor results getting articles published in the past, but this is something I want to get out of my system, so I am giving it one last shot. I don't intend to attempt to formally publish articles after this. I think it is more productive to hash out ideas out in the open on the net. For instance there was a lot of constructive back and fourth about list performance in scripting languages. You just don't get that from reading a magazine in the bathroom.
This will also probably be the last C++ project I will embark on for awhile, outside of maintaining SwitchFlow. C++ has been good to me. It landed me a scholarship, my first jobs, and provided a nice living over the years. I still love the power of the language, but realistically it is becoming more of niche language, and I think the C++ community is driving it that way as well. Virtual machines is where it is all going IMHO.
IronPython on Mono represents the current state of the art of OpenSource development tools. I am amazed at the work Miguel and his team has done with Mono, and I'm happy Microsoft hasn't done anything to sink Mono, IronPython, or IronPython on Mono for that matter. I'm starting to have a different opinion of Mircosoft these days, but I'll leave that discussion for another day.