« Narrative Accommodation and Gameplay Growth | Main | In case you were wondering... »

September 16, 2010

Wasted Features

One of the last things I wanted to briefly discuss regarding "my divorce" was a wasted feature I included.

What makes a feature a waste? To me, a wasted feature is one in which there is an unfavorable ratio of "number of people using that feature" against "time taken to implement that feature".

Wasted features might also represent features that weren't fully explored or exploited (which was more or less the subject of Chris Hecker's rant this year at GDC). Keeping that in mind, it might be better to reframe that ratio with "enjoyment hours" as the numerator. Measuring the amount of enjoyment from a particular feature would be extremely difficult, but that's often what we're trying to deliver. Features that are more fully explored are more likely to give people pleasure, or give people more pleasure, so it’s worth exploring your mechanics in depth as far as you can think to take them in the time you have.

This is a highly utilitarian measure - we pursue those things that give a decent bang for buck. After all, developer time spent on something few people enjoy is a lost opportunity; that time would have been more wisely spent elsewhere¹.

In any case, the feature that I feel was a real failure in the case of "my divorce" is the "copy-and-paste-to-share" feature. The game supports the ability to hit the standard Windows copy key combo, and will put a compressed version of the game's data in the clipboard. The thinking behind the feature was to enable a Spore-like sharing of data via a lightweight format (a few hundred bytes, say).

It's a silly feature for an art-game to have, but I was trying to allow users to specify their own views of their own divorces, real or imagined. The simulation is highly configurable via an XML file -- you can make any number of statements through the data in combination with the rules. You could make different events impact the “children” in different ways, you could represent divorces with same gender parents, or say that gender doesn’t matter. A lot is possible, and so I wanted a simply way to allow people to share their own data, and an XML file just seems so inelegant.

I doubt the feature has gotten much use, if any, and so from that perspective it scores very low on our utility function. On the other hand, I really only built “my divorce” out of my own need to do so. In that sense it only needed to satisfy me, and it was a fun little bit of code to write. On the third hand, the code is very generic, so it would be very easy to repurpose in a future game -- perhaps as a means of sharing level data. So some day its utility may be better realized.

The underlying assumption -- that time spent on this was time not spent on other more beneficial features also doesn’t really hold. The game is exactly what I intended, and I released it only when I felt it was completely finished.

So, this feature was wasted in my art-game, since I suspect the amount of enjoyment it gave to others (other than myself) was effectively zero, but hopefully it will find other use in the future.

¹It's possible to get very meta about this. For example, I really like the sense that I can take many paths through Deus Ex, but the fact is I've taken only one. But I recognize its depth of simulation as a major contributor to my enjoyment of the game.

Posted by Brett Douville at September 16, 2010 08:42 PM