Considerations with a site re-design

I just completed a site redesign for the Alberta Roofing Contractors Association. They wanted to refresh the pages a bit, simplify the navigation and add Google Custom Search. Here's some things I learned...

We switched from static html pages to php pages that used simple includes for the header, footer, and navigation. We also reorganized the folder structure a bit to clean things up. Consequently all the pages/pdf's that were previously indexed on search engines 'disappeared'. I wasn't sure what to do about this at first, but then I found an .htaccess file in the root of the ftp directory on the server and I made 301 redirects for all the old html pages and pdf files. This worked great. Lesson Learned: a simple site redesign on an established site isn't as simple as it would seem on first glance.

I found re-indexing Google custom search to be a pain. I had to add the site into Google's webmaster tools (both www.arcaonline.ca and arcaonline.ca), set the old pages up to be removed from the cache, create a sitemap.xml file and upload it, create a robots.txt file and upload it, and add all the new url's for pages and pdf's in the 'explicit' indexer. It was rather cumbersome, but I got it to work in the end.

There were a number of issues with compatibility related to the search engine form and the flash on the home page that we had to work around using CSS hacks for IE ( *+html ).

Using a wrapper in CSS is a big help in some cases to make background images display properly in all situations. Something similar to this:
#wrapper {
margin: 0px;
padding: 0px;
width: 100%;
background-image: url(../images/bg.jpg);
background-repeat: no-repeat;
background-position: center top;
margin-left: auto;
margin-right: auto;
}

Automatic Deployments taken to a new level

I was doing some research for an article I was thinking about writing on Automated Deployment and came across this fantastic blog post by Grig Gheorghiu. He talks about how OpenX uses Amazon EC2 (virtual machine cloud) to automatically provision and terminate servers of different types, different monitoring solutions they use, and in another post he talks about the different automated deployment technologies that can be used (push vs. pull, ruby, python, etc).
I personally liked the comments about 'never having to log onto a production server', and "There's nothing like real-world network traffic, and I mean massive traffic -- we're talking hundreds of millions of hits/day -- to exercise your carefully crafted system infrastructure."
All the comments at the bottom are really good as well.