A couple of years ago I built a small desktop app to download 80,000 (or was it more?) images from Flickr. as part of a massive Birmingham photo project called The Big Picture. Today, as I sit watching a terminal window list the URLs of photos it's downloading from the same site, I wondered what would have been different if that app had been written in Python?

For one, it wouldn't have needed a GUI, or an installer (bat that needed to get Python running, which is really simple). The user would simply enter the right credentials, the tag(s) they wanted to pick up, and the directory they wanted the photos to be dumped into. Then just sit back, and watch the terminal window spit out the URLs, so you know it's working.

Maybe the one thing that did speed up the process in the original C# .NET application was the threading. If I recall correctly, the app could download two or three images at a time, or at least download them one at a time and on one thread, while the other thread handled the progress bars. But threading in Python is fairly simple, so it's possible something could have been knocked up, although how useful it would have been is another matter.

Because Python is a scripting language, and doesn't need any manual compiling - unlike the C# app I built - fixing issues and distributing those fixes would have been simple, and it could even have been run as a web application, using mod_python or a WSGI handler in Apache. Plus, there is already an extensive Flickr API wrapper for Python, so there wouldn't even have been that much code to write.

As it happens, I didn't start to learn Python until much later in the year, and it wasn't long after that that I ditched .NET altogether in favour of what I saw as a much more readable, flexible and beautiful language.

In short, if I had built my part of the Big Picture in Python, the actual coding would probably have taken a couple of hours, as opposed to the day or so it took me to build a pretty-looking - and possibly slower - desktop version in C#.