On Jan 26, 2009, at 10:44 AM, Russell Wilson wrote:
And how do you objectively prove it's better? Short of testing alternatives you have a very subjective problem.
It is possible to objectively test any design problem, as long as one remembers that the definition of the problem is largely subjective, or at least arbitrary. You have to measure against the definition of the problem, not some arbitrary set of alternatives for the sake of A/B testing or whatnot.
Further, there is rarely if ever "one" perfect design solution, but many very good solutions possible; any and all of which work well enough within reason and to varying degrees. This makes it more important to define your ranges and criteria of a problem and measure against that. "Objectively" is relative then to how well you've defined the problem in the first place.
Think of it this way: Is there only one man or woman for you in the world to fall in love with? If you believe so, good luck finding that person out of the 8+ billion on the planet. Most who try live a life of constantly striving for some ideal of perfection that forces standards on their companions which are both unfair and often untenable tend to be disappointed. However, if you approach the problem reasonably, you'll know there's a range of possible people as defined by yourself. Once you find someone in that range, you can date (or do whatever it is you do) to test it out then fall in love and become soulmates forever, and even then, you have to evolve with each other year after year.
Finding the right design for any problem is often the same kind of task. You "objectively prove" something is better by measuring against a variety of agreed criteria, and analyze the results based on tangible data. Things like: Less line noise, a request to use fewer colors in the display palette, fewer keystrokes required for task A, solving the problem of allowing 12 tabs on screen simultaneously with text strings of 30 or fewer characters in them, working in an interaction model similar to something else in the target user's mental model so that keyboard shortcuts match what they are used to, customer testing that yields positive feedback, proof of shorter development cycles from engineering, etc. The criteria is completely arbitrary by the way but always specific, defined by someone on the team or the team collectively. Once you find a design that works with the criteria -- not the perfect design, just one that works by some set of standards -- you can work towards making it yours, own it for all it's pros and cons, and evolve with it as time moves on.
There's nothing strictly "objective" about it until you have defined the problem with some level of specificity. And measuring against it is often a range of pros and cons. The definition and criteria of the problem is all that matters. And I stipulate a lot of the problem between designers and engineers is that they both tend to define the "problem" differently. If you have a team all treating the problem in different ways, the first thing that you do as a designer or team lead is baseline the problem so everyone understands the definition or a new definition is built that covers everyone's criteria. If you don't, what you wind up having is a discussion based on some people liking blondes and others preferring brunettes, and both treating the issue as if it's a religious discussion, which leads to bickering and arguing over useless details that will never see resolution.
-- Andrei Herasimchuk Principal, Involution Studios innovating the digital world e. and...@involutionstudios.com c. +1 408 306 6422 ________________________________________________________________ Welcome to the Interaction Design Association (IxDA)! To post to this list ....... disc...@ixda.org Unsubscribe ................ http://www.ixda.org/unsubscribe List Guidelines ............ http://www.ixda.org/guidelines List Help .................. http://www.ixda.org/help
