Tomorrow, we're heading into the hospital so my wife can have an amnio to verify that the lungs are developed enough for delivery. If they are, she'll be having a C-Section Friday. If not, it'll be Monday. She'll be 37 weeks and 5 days on Friday. Most babies lungs fully develop by 36 weeks, so most likely our second daughter will be an April Fool's baby.

It's an odd feeling knowing when it's going to happen. I remember with my first daughter when my wife told me she was in labor. It was 3:30 AM when she woke me up. She took a shower, called the doctor, and we went in. 26 hours of labor later, we had our first daughter, by C-Section. I was exhausted - I can only imagine how she felt!

This time though, it's at 3:30 in the afternoon. She'll get up early, have a light breakfast, then get some more sleep. She'll be sore after the surgery, but she won't be exhausted this time. It should make things go much more smoothly. I hope.

Anyway, I'll be at the hospital probably Friday through Monday or Tuesday. That means I won't be posting much, so expect the blog to be silent for a few days.

And of course, when I get back, I'll be sleep deprived, so there's no gaurantee that posts next week will be coherent anyway!

Craig Andera has a nice post about building applications, pointing out that you're never building just one.

He says that while the goal is to build an application, you have to build auxilary apps to go with it. You need to have tests, a build process, and a way to deploy the application.

I was reminded of this today. We went live today (I'm waiting for the official word to make it's rounds before I let the cat out of the bag on what I've been working on), so today was spent revamping our build process. We use Nant to do our builds, usually through CruiseControl. But when we want a to do a release build to our test environment, it's done using just Nant. To get us by while in crunch mode, we got something up and running, but it wasn't optimal.

With the site live, it was time to fix that. So today, I spent some time in our build files. The end result? First, I got intellisense working in the build files, using this post from Peter Marshall. Now, can you get both nant intellisense and nantcontrib intellisense at the same time?

Next, when we want to build to production, it's as simple as running one command. That command does the following steps:

• Stops the CruiseControl.NET service (this prevents builds from interfering with our prod build)
• Cleans both the dev and test directories
• Gets the latest into the dev directories
• Builds in release mode
• Copies files into the test directories
• Zips up the builds so they can be copied up to the production servers
• Optionally, it labels VSS with a command line specified label.
• Starts the CruiseControl.NET service again.

I also made it so each developer can have their own properties file that defines certain properties needed by Nant to do the build. What does that do? Now, every developer can run Nant locally to do the build, something we didn't have until now.

I also added the ability to specify a label to build. So I can specify a label two or three versions ago, and be able to build that automatically. Now, we can go back and build our Alpha build from five months ago, and see how far we've come since then.

Overall, I'm pretty happy with the build as it stands right now. If anyone wants to see a sample build file doing the things listed above, let me know. I can post that if there's enough interest.

A few days ago, Yahoo released their first beta of thier blogging service: Yahoo 360. Tom over at WebFeedCentral was kind enough to send me an invite, and I signed up. Thanks Tom!

So far, I'm not terribly impressed. To even view the blog, you have to be a Yahoo member. To comment, you have to be a Yahoo 360 member. Both of those seem backwards to me. So I can only get comments from people who blog at Yahoo? Odd. Maybe it's because it's still in beta. Maybe it's their way of getting people to sign up (of course, you can't sign up directly right now). Who knows. But the fact that no one is really answering these questions seems like an unwise decision.

My initial impression says it's more than just a blog though. It's kind of a mix of a social networking software (think Orkut) with a blog thrown in. And Blasts.

I haven't been able to see an RSS feed from Yahoo 360 yet, but word is they are adding a copyright notice to your content, but it looks like that's a mistake.

Anyway, if you're interested in checking it out for yourself, let me know. I have 100 invites. And if you want to see mine, it's available here.

Fanblogs has a post about a new bill in Texas that would bar Texas teams from participating in BCS bowl games. Basically, the bill says teams have to participate in a playoff system to be eligible for post season play.

Of course, it's mainly just for show. It expires before the BCS bids go out unless four other states jump on board. But not just any four states. It's from a list, and the list pretty much includes the states that would actually send a team to a BCS game.

Personally, I'd love to see a 16 team playoff in college football. It would be as big as March Madness! And I'm not just saying that because it would be supported in the Tourney Pool Manager. It would be great for the game. I think last year, when USC had a legit claim on the national title, it was pretty obvious the BCS wasn't producing the results that it was put in place to do.

What about bowl games you ask? Well, use them to host the playoff games. With 16 teams, you can incorporate 15 bowls into the system. Yes, right now, we have 497 bowl games, but cutting it to 15 games seems about right to me.

Hopefully this won't be a token bill. I think it will be, but maybe it'll get the process started.

Yesterday I posted about a blogging expert who said he doesn't see the point of RSS. Well, today he visited my site, and commented, saying I fell for his bait. He was looking for someone to "expose the truth." His comments were pretty nice, considering my post was ripping into his position.

I remembered where I found the original link. It was from the Hello World Blog. And Steven left a comment there too, but he wasn't as nice over there. He called them losers.

Then I went to visit his blog. He's changed his tune on syndication. He now offers a feed link on the site, which is good.

But he removed his post about RSS, and takes his "enemies" to task for disparaging him. He questions us - what stands have we taken that aren't "safe"? There's a difference though. We're not claiming to be blogging experts. He is. And he said he didn't see the point of RSS, and now he's removing posts. And he says it's a weakness that he's admitting he's wrong. It's not weak. It's actually a strength to be able to admit when you're wrong.

At least he changed his stance on syndication.

As part of getting ready for deployment, I cleaned up our sourcesafe database, removing files that were at one point part of our project but no longer are. This allows our automated build process to pick up just the files that should be there. We use Nant and the solution task, so we don't get files compiled in that shouldn't be there, but we do get extra files in our directory, so it was time to do some clean up work.

Anyway, as I was removing files, I noticed a lot of the files that got removed weren't really removed. They were renamed in VS.NET, which causes an extra file to be added to VSS. Seeing this, I was reminded of the correct way to rename files when using VS.NET and VSS, so here it is:

1. Exclude the file to be renamed from the project. You'll have to check out the project to do this.
2. Rename the file in VSS.
3. Get the latest of the renamed file, and add that file to your project.
4. Check in the project file.

Doing this prevents erroneous files in VSS, and it preserves the file history. Both very good things!

Remember though, that when you do delete a file from VSS, don't destroy it permanently. That way, it's there if you need it, but not there if you don't.

As we near the final days of testing, we're also getting ready for deployment to production. That means all those configuration settings need to be looked at, and adjusted to be production ready.

One that I always see is to change the compilation tag in the web.config to have debug="false". It makes sense - you don't want debug code in production. Combine that with a release build, and you'll get the better performance.

So we did that. I tested it locally, and immediated had issues. It said a user control was defined in multiple places. I couldn't figure anything out, but the same control name was used in multiple places, just in different namespaces. I solved the issue by renaming one of the controls, and the web server was happy again. But I don't know why that worked. We have other user controls with the same name, but that one caused problems. Very strange. Any one have any ideas?

Regardless, we are in the final hours before we go to production. I'm excited. The client asked me if I was nervous. I have a pretty big project that I was the technical lead on that is about to launch, there's the Final Four, which Tourney Logic has a big stake in, and then I have a baby being born. All finishing up within a week. And my wife is on bedrest. You'd think I would be stressed. Oddly, I'm not. I'm just ready to get them all out the door, so to speak!

I forget where I saw this, but someone pointed out Steven Streight's blog, where he posts about why he doesn't like RSS. After reading the post, it's pretty obvious it's not that he doesn't like RSS. It's that he doesn't get RSS.

He says RSS pushses content at me. It doesn't though. I subscribe to a blog, and then I pull the content when I want it. He compares it to IM, which he doesn't use either. But you don't ask someone if they have something to say to you over IM. That's what RSS is. You ask the site if it's been updated, and if it has, it tells you. It's not unwanted. I hope he doesn't use email either. Now email is unwanted pushing of content on you!

He goes into a few more reasons he doesn't like RSS. The first one amazes me. He says he has so many bookmarks that he can't manage them. I can manage mine, but only because of RSS. I don't have to bookmark sites. I subscribe, and then I know when the site is updated. I don't have to visit the site. The site comes to me. His second reason is the same as his first - he has too many sites to visit each day. That's because it's slow to visit sites. RSS lets me grab the content automatically and read it when I'm ready. I wonder how many sites he monitors every day. 15? 20? 50? How about 500+? That's how many I monitor. But only because of RSS.

I think Scoble has commented on this same thing. I can't find a link right now, but usually about once a month or so, he says the same thing. And he monitors 1300 sites a day. Let's see someone do that in a browser!

The best part? He's in the process of selling a book called "Secrets of the Blogging Pros" which is a book about blogging's best practices.

Tim Haines says you shouldn't develop in the web root. And he gives some advice to get around some of the gotchas with doing it.

I agree with Tim. But here's how I set things up. My main development happens in a virtual folder on port 80, as does most of my debugging work. But I also have another website (on another port, like 87 or 88) that has it's root as the same folder as the virtual folder on port 80. That way, I can test that everything works correctly when the app runs as the root.

Other things I do? I have www.localhost.com mapped to 127.0.0.1. Why? We ran into a few URL parsing issues that didn't work correctly when there were periods in the domain name. Of course, I'm missing out on whatever www.localhost.com really is!

All of these issues should be something taken into account when developing the software, but in big teams, inevitably these issues will pop up from time to time. Better to have a way to develop and debug in all the cases locally than try to debug a problem on a remote server.

One more post about religious wars - this time about Keyboards versus Mice. Peter Provost agrees with me. The keyboard is the way to go. I know the keyboard shortcut for just about everything I do in VS.NET. And if it doesn't have a keyboard shortcut, I'll make one. For example I undock all of my windows, and use the keyboard commands to show them. The Pending Checkins window doesn't have a command, so I mapped it - Ctrl-P, Ctrl-C (Pending Checkins - makes sense to me!).

I use a laptop, and having the little ball on my keyboard to move the most means I can still use the mouse without my hands leaving the keyboard, but I still don't use it. And I never connect an external mouse to it!

Don't agree with a complete keyboard solution? Peter points to aboutonehandtyping.com, which is a site dedicated to typing one handed, so you can always keep your hand on the mouse. I guess some people are just as passionate about using the mouse as I am about not using it!