For our app, we went with the iframe option, because we had lots of data exchanges going on with our own database, and users needed to upload pictures as part of using the app. During testing, (in various flavors of IE, Firefox, and Safari on both major platforms), no real problems were encountered with the functionality of the app that weren’t resolved before deploying. Once deployed, however, some users were encountering problems that seemed to be random glitches that we couldn’t really troubleshoot.  The API methods for getting a user’s photos sometimes didn’t return anything, no errors, no photos, nothing.  This actually seems to be a known problem, and an obviously annoying one.  Also, sometimes the user would be taken outside of the iframe, and directly to our server, which is obviously not ideal.  This was fixed with a kludgy workaround, where all links were pointed to the app’s URL on Facebook, with our URL parameters appended.  This, however, is going to make tracking less accurate with Google Analytics, since they are technically not links within the iframe any longer.

So I’m curious, has anybody else had problems when developing Facebook apps?  Does it feel as messy to you as it did to me?  It’s been an interesting experience, and I just hope I’m not the only one that feels this way after my first real Facebook app.

In related news, I’m finally getting to the point where I’m actually redesigning my other blog, so stay tuned for posts on the entire process of re-creating that site.

Cheers.

This doesn’t mean I’m leaving PHP in the dust – I still use it extensively at work, with no plans of stopping anytime soon – but I just have a feeling this is a language I would like to work in. Python’s syntax is similar to Ruby’s (at least at first glance), and from what I’ve read it outperforms Rails by quite a bit.

The project that I’m going to cut my Django teeth on is my wife and I’s blog: www.rustyandben.com. I talked briefly about overhauling the site a few months ago, but some other higher priority projects came up. I’m getting to the point now, however, where I can start working on it again. I’ll be re-designing and developing from scratch (obviously maintaining all data), so it makes the perfect opportunity to try out Django. I’ll be upgrading my Media Temple hosting to a virtual dedicated server so that I can have root-level access in order to install whatever software I want. This, in turn, will allow me to try out a couple other technologies I’ve been wanting to check out: SQLite (a lightweight flat-file database system) and Lighttpd (a lightweight and fast HTTP server). More on those as I assess their feasibility.

I’ll be posting about all this again soon, as I’m planning on writing a series of posts about the creation of the site.

Some other things that were of concern to me were my various web development extensions – Firebug, Web Dev. toolbar, FireFTP, Dummy Lipsum, etc.  This problem was solved by installing Oxymoronical’s Nightly Tester Tools. The extension allows you to disable compatibility checking. Then you can enable your extensions one by one and make sure they don’t break anything. Luckily, all of my must-have-extensions work perfectly with compatibility checking turned off. Yay!

Finally, I recommend using the Firefox Nightly Builds, as opposed to the betas, so you can get the latest updates (and maybe a bug or two). Not a big deal though, and it’ll update itself to stay current.

There you go, fully functioning bleeding-edge install of Firefox 3, with (hopefully) most of your extensions working! By the way, it installs itself as “Minefield”, rather than Firefox, because, well, you’re taking a chance by installing the nightly builds .

Enjoy.