MSN World Tour // Post Mortem

I've just rolled off the MSN World Tour project co-produced by Wunderman and Navantis. I was contracted by Navantis to lead the Flash team, integrate Flash with .NET, and write the client-side game logic.

What Went Wrong:

Unless you've produced a project like this before, the planning is next to impossible. On the surface, it seems like a straightforward contest but there all kinds of technical obstacles once we dove in and began deeper delevelopment. For instance, the flip book -- as cool as it is -- was complete headache to hack into the game and make it behave properly during different conditions such as when the user closes their browser and we have to recover that state when they return opening to the proper page. It's also partially written in Hungarian! If you are interested in developing this trendy flip book for your next project, check out the open source source.

MSN security demands for the user encryption during transport were also extremely challenging in Flash [ you can read more about that here ] and relying on external Javascript classes and primitive Flash 4 fscommands.







The code base for this game was massive in the end. The Flash client itself has well over 10 thousand lines of code and this became increasily difficult to manage between two developers, and a designer who actually contributed quite a lot of code too -- well done Mike!

There are few minor things that I would have liked to have in the game but were tossed due to time concerns and were not essential to the experience such a keyboard support -- I know this isn't great for accessibility but hopefully something that the next team can build in as this contest will probably be ported to more than 8 different languages. We did both French and English. Finding French translations for everything was fairly troublesome, requiring more than 24 hrs turn around from the translators -- so yes we used babel fish occasionaly to save some time on very minor user messages.

The main business goal of this contest for users to install the MSN Search Toolbar. MSN provided javascript code to detech the presence of the toolbar, however, it was not relied so we had to a survey in the registration if the javascript failed.

Personally, I would have loved to see the search results for the hints displayed in Flash instead of jumping back and forth between browsers but this went against business goals to drive users to the html search pages.

What went right:

I won what could go down as a legendary IT battle to keep the registration process in Flash! The major selling point was that we could save time by re-using the pre-registration and create a more seamless user experience.

I was a little worried in the first week when Navantis opted not to invest in a Flash Remoting server, but the sendAndLoad() mechanism in Flash worked out very well. The lead .NET developer at Navantis, Reuben Diel, produced an extremely helpfu xml interface document which we iterated and updated throughout the development cycle as we discovered missing elements.

Our project manager, Adrian Celentano, did a great job of distracting the client and shielding the development effort.

And a special thanks to our one-man calavry Taylan Pine who joined the Flash development efforts late in the game and really helped to clean up any missing elements like eastern eggs, winks, and the kitchen sink!

So...all things considered for a development view, this was a successful Flash /.NET project, and I'm sure it will pay off in the long run for Navantis and Wunderman -- its expected to reach more than 10 million users!