This is one of those “duh” type moments, but I just realized that, even though SP2 installs HTTP.SYS, it stil uses IIS 5.1, which means IIS is NOT using HTTP.SYS.  This would explain my recent trouble and why trying to control port usage with the httpcfg tool didn't seem to do anything.

I'm pretty bummed about this. I probably would have installed 2003 server instead of XP had I realized this sooner.  I'm going to give a serious effort to write a managed Subversion server that uses HTTP.SYS.  That would probably reduce the barriers for using Subversion sufficiently to get more people interested in it.

[UPDATE]  In case you didn't make the leap and realize what the implications are for IIS not using HTTP.SYS, it means you can listen on the same port that IIS is.

I don't really want to become a link blog, but Brian Grunkemeyer has an exceedingly interesting piece on arrays and how they mesh with generics in the 2.0 CLR.  He goes into alot of interesting information that you don't need to know, but knowing makes you appreciate the the work going into it.  I've had the pleasure of speaking with Brian, and thoroughly enjoyed it.

Pardon my ripping off of the title, but I really found that it quite succinctly summarized the post.  Don Box has a pretty valueable little tidbit on doing just that.  Very handy indeed.

I'll have to digest it a bit to see how applicable it is to 1.x.

Today is my little brother's birthday!  Happy birthday Andrew!

OK, I was not prepared for the onslaught of hilarity that ensued in the comments of my last post.  Somehow I think it's not over yet.  It's good to know that even though there are alot of us Mark Millers, that most of us seem to have a good sense of humor.

I about fell out of my chair at the thought of someone stirring a butter knife around in the trash to make it a more damaging weapon.  I think I'll have to use that in a movie.

For a split second, I was baffled at Scott Hanselman's recent post which claimed I had been working with him on some documentation tools.  Then I realized that, of course, this was a different Mark Miller.  This was not the first time I've been frustrated by the seemingly large number of Mark Miller's, especially in the blogosphere.

The real question is how can I distinguish myself from the others?  Certainly not by continuing to post rather bland entries to my blog.  I'm not able to make my work blog public (which is what I'd like to do), where there's alot of nifty activity.  I guess I just need to make a more concentrated effort to get my thoughts out in the open without revealing the application of the thoughts.  This probably could have a nice side-effect of thinking a little more abstractly about the decisions I make and maybe see a more broad application than I originally had in mind that make make a design work better.

Also, why WASN'T I working with Scott on some project? Why aren't I connecting with others more? That's a problem.

So, stay tuned, i guess.

I finally got Subversion set up over SSL hosted by Apache side by side with IIS.  The biggest problem was getting IIS to let go of port 443.  I'm running XP Pro SP2 and I'm not hosting any SSL from IIS, but it would still bind to port 443.  I found lots of information on configuring HTTP.SYS using httpcfg.exe (There's lots of other very cool SP2 tools in the download), which I though was really cool, but nothing that kept it from binding to 443.

My rather rudimentary solution was to start Apache first.  IIS/HTTP.SYS didn't complain about the port already being bound to.  It works like a charm now.  Now I can get to my repository from work through the firewall!  What I don't know is if the sequence of events at boot time will keep it from working right.  Any ideas?

I'm wondering how hard it would be to write a Subversion server on HTTP.SYS.  Now THAT would be cool!  I think it would only be a matter of implementing (or finding an implementation of) WebDAV and hooking into the Subversion API.

I am not a morning person.  Anyone who's ever spent any time with me in the morning knows that.  I've been tracking my productivity more closely lately (in terms of tangible, work-related results anyway), and I've found that I am far more productive after lunch than I am before lunch.  This doesn't mean I don't do anything in the morning, I just don't seem to have the same kinds of satisfaction with how I've spent my time.

I originally thought that I was being productive in a different way in the morning.  For instance, in the late afternoon and evening, I enter a fairly reflective state.  In this state, I question the design decisions I've made or the implementation I've chosen.  This is when I do alot of measurement in the form of tests or profiling.  That's plenty productive and useful activity, but in a different way.  I thought maybe the morning was perhaps an analogous, forward-looking phase, readying me for the tangible productivity ahead.  But, no, I'm just not very productive in the morning.

I'd like to find some activity or state of mind I could concentrate on in the morning in order to accomplish more in the morning.  Any ideas?  Does anyone else feel like their mind is more dormant at points during the day?  What do you do about it?

This morning I looked up my voting location and found it to be less than 100 yards from my house!  It was quite convenient and I had voted in a matter of minutes after a short walk down to the community center.  I'm glad I was in town today and wasn't forced to wait in a long line for the convenience of early voting.

I though this article was appropriate for today.

I hate the fall time change.  When I was young, I was always was fooled into liking it because I got an extra hour of sleep.  Now I know it's just a trick to take one hour of daylight from the end of the day.  I can get an extra hour of sleep pretty much whenever I want by going to bed earlier. I'd much rather drive to work in the dark than come home in the dark.  It makes me just want to go to bed when I get home, and it doesn't really help me wake up any earlier.