Designing with rules

If there’s one thing I wish I could tell all immersive designers getting involved in immersive media and interactive technology is that they’ll be far more successful if they created rules instead of designs. Unfortunately, real-time systems aren’t magic. Immersive installations are built on rules. Sometimes many of them in complex chains that create elaborate decision making trees hidden to users. But these rules don’t just appear on their own, they’re created with intent by designers and developers.

What do I mean by rules?

Absolute definitions of the why behind your designs. You’ve probably talked through designs in your head, deliberated on decisions until you were happy. You’ve iterated with different text revisions sent to you and different sized logos and each time you had to nudge things a certain way. At some point in time, you knew exactly why things in the design turned out the way they did. But then these decisions are quickly forgotten as beautiful designs are rendered to be presented up/down the chain of command for approval. The problem with this is that most immersive installations rely on data and data is an always changing beast.

Data sucks (most of the time)

The render of a single data set in a design won’t cut it when you start to scrutinize all the individual elements at their possible extremes. Examples:

• What you thought looked great as a title with 50 characters might blow off the screen at 150 characters. What then? Do you resize the text dynamically to make it fit? Should you newline it? What if it’s a single word, should you hyphenate it?
• Maybe the layout you’ve designed to split up 150 names into separate columns looks really balanced and fills up the screen until there are less than 80 names. At that point do you start just removing names from the last column and leave white space on that side of the composition? Do you instead create an even split of names for each columns but then leave more white space below the columns? Do you resize the font to take advantage of more space? What if you had double the amount of names all of a sudden? Would you create 2 “pages” of columns and fade one out and fade in the other? Would you create a scrolling effect to scroll the columns across? Would you make the columns smaller so that you can fit more columns on the screen?
• The stock market bar graph you designed looks great when all the values are between 15000 and 20000 (which they have been for maybe 10 years!), but what happens when the value escapes those ranges? Or worse, what happens when the API returns some false data (as all API’s do eventually)? Do you dynamically rescale the intervals on the chart axis so that you can fit all the data within the same region on the screen? Do you increase the graph size and allow the value to leave the graph dramatically? Do you add a break in the bar of the values out of range so that you can keep the chart the same size and avoid dynamic interval changes?

Data is unpredictable and uncontrollable.

Strategies to win

When I’m working with designers who might not have much experience in interactive and immersive installations, the first thing I do is get them to talk me through their decisions. I ask them why things are the way they are and if they’ve looked at the same design with different data, which often results in all of their forgotten design decisions coming back instantly. I try to talk them through may different angles and warm them up to the idea that the rules are more important than the designs themselves. My goals are to get the designers to do 2 things:

1. Write down all the rules so they can easily be read through
2. Mockup some edge cases

Writing the rules down allows the rules to get reviewed and approved alongside the actual designs. This is critical!!!!!!!!!!!!!!!!!!!!! I don’t know how many times I’ve been in situations where designs were approved, a new data set gets input just before the client shows up, and then the client asks why it looks different than the designs…..facepalm. If the rules are written out simply, then the same people approving the designs also get a sense that it’s a living thing and know that the rules will allow it to respond in a way that they like.

Mocking up edge cases doesn’t have to be a whole ordeal. I understand renders of complex scenes can take a long time, but in most cases you may not need to actually render anything. If you can see your designs, change the info, and have low res previews in the software, that’s often more than enough to give the designers a sense of whether the design holds up in extremes.

Final trick

Use real and live/changing data as soon as physically possible. This sounds obvious to say but it’s hilariously always such a low priority in the hustle and bustle of pure aesthetics, stability, client approval, installation timelines, budgeting, staffing, hardware procurement, etc etc. The sooner you can get even a rough version of the designs feeding off live data and demonstrating to the designers that it’s a live system, they’ll immediately become aware and quickly start to reconfigure their design approach to match the unpredictable data.