A few years ago, as part of a contract we were servicing, I was asked to take over Webmaster responsibilities for a suite of open-source Web applications we were developing. An important component of the project was its open, or public, nature - users were invited to participate directly in the development process by providing feedback on development snapshots released by the development team to the Web site on a daily basis. Since our customers for the software were located in a different country, these daily development snapshots also provided them with an easy way to check on the progress of the project at any time.

As Webmaster for the project Web site, it became my responsibility to ensure that the site was always running the latest build of the software, so that users could play with it and give the development team feedback on how well it was (or wasn't) working. In the beginning, the task was easy - but as the project size grew, I found myself spending more and more time in front of my FTP client, watching as one file after another slowly wended its way from our staging server to our Web host.

Now, you have to keep in mind that, at this time, no one had heard of broadband, and so most of these uploads took place over a slow modem link. Since I had no way of knowing which files had changed between builds, I usually just uploaded the entire source tree from our local servers to the Web server - a long process, and one which grew ever longer as the project matured and the development team added new features. What I *really* needed was a way to just transfer the delta - the changes between the last build and the current one - so as to reduce both the time spent by me on the task, and the cost to the company in terms of connectivity charges.

That was when I discovered rsync.