1. Stop making bottleneck software
Technically you could write most software today as one big mega-class with loads of switch / if&else statements. If you did that, not only would every other developer you come across immediately punch you in the nose but it would also become hard to maintain over time.
We agree that would be stupid right. I mean one large file for all code! - yet why do I always see software designed in such a way that it becomes the Swiss army knife of all tasks associated to the user, in that it becomes feature-heavy based around feeble arguments of "but the user wanted.."
The user is 80% of the time a jackass.
You are armed with a plethora of programming models today, stop crowding (thereby creating UX bottlenecks) the user interface for every single role known to man. Figure out the "persona's" attached to your software and if need be, make smaller contextually relevant versions of the software per person (whether it be modular or separate specific installations).
2. Third Party Controls do not negate the need for a designer.
When I first left Microsoft and joined the working class (mwahah), I was often thrown into the deepened of projects that needed some UX makeovers. Given I have both a programming and design background it seemed a natural fit so sure, go with the flow I say. I'd walk into a typical gig and sure enough I'd see 3rd party controls lurking about (ie Telerik, Component One etc).
Nothing against these brands but if you are dealing with WPF or Silverlight then let me give you a heads up on why this is a bad idea. Firstly, the 3rd party controls are just a quick dirty fix to get around bad UI design, I get it, budgets are non-existent so you do the best you can. Secondly, these controls are made for multiple developers around the world, so there are many keys to turn on and off for them to snap together - which means your controls are not on a diet. Thirdly, you need to walk a mile in the shoes of say a C++ programmer or some language that used to have to play a game of memory Tetris to really grasp the concept of the second point.
Diet is the keyword. If you are dealing in Silverlight space the leaner and smaller footprint your code has, the snappier things are going to get. I am not talking about pure CPU no-holds bar processing time; I am talking about rendering pipeline time. I am yet to see an example of 3rd party controls improving performance and not subtracting them.
Stop outsourcing design for third party controls and I am looking at you graph boy/girl.
3. Every screen has a soul.
In UI Principles space there's this little concept call false affordance. It means something that looks like it was supposed to do xyz but does nothing (i.e. Push the button and all negative energy will disappear scams).
If you have some software that has a hierarchy of navigational elements, you click on the first node, and it does nothing but expand to the second node but at the same time shows a view with some "weak" summary (i.e. description etc.). Stop, you are doing it wrong.
Every click has a purpose of existence. If you have a dashboard, what is its purpose? Think about its relevance in the grand scheme of things. Should it be fresh content daily/weekly/monthly? Is a holding pattern screen necessary? The screen, which is like the UX principles are buffering between two major waypoints - you know the one screen in the app that really has no purpose other than to get you from A to C but somehow you felt the need to keep B in place.
If you have a screen that is filled with say two Input controls and that is it. That is a freakin dialog box, it is not a screen. Stop being lazy and think about the problem not how easy it is for you to just whack up an app. It's not about you, it is about them *points to the end user*.
4. You are not a magician so quit giving me the constant "surprise" moments.
Ever used an application that when you click on something random inside a screen suddenly a piece of User Interface randomly appears somewhere in the screen? Maybe hidden inside a secondary tree node somewhere?
This is not a magic show and you are not a magician. Progressive Disclosure is great when done in a way that leads the user on a journey, no more "I've just modified the screen, if you guess what I just did you get a fluffy kitten" moments.
5. Humans are smarter than you think
I have covered this quite a lot but let me re-iterate in the theme of this post. Over 90% of the world's computer population right now has some piece of overly complicated unnecessary piece of crap software installed on their hard drive that they have somehow managed to figure out partially its inner workings.
The benchmark for success right now in this space is so low you could trip over it and still succeed. My point is the end users are actually smarter than you give them credit for. If you are in a team and someone says, "Yeah our users aren't smart enough to.." challenge that jackass upfront. As did he conduct a survey where One in Five housewives came back being dumber than he anticipated?
If an average worker-bee can sit through SAP ERP or any piece of software that Oracle/Microsoft throws at them, they can sit through your software as well.
The trick is to make it enjoyable for them, to be the software that does not feel like the others - the stand out. Rather than holding them hostage to complexity because of your own arrogance, try to think less about the complexity levels and more the enjoyment levels. Software should be enjoyable as we work WITH software - we do not USE it.
6. I did not buy a cat so it could be my master.
My kids wanted a kitten and so me being the "fun" dad I bought one. Today that cat rules the house most of the time because we react to it, not the other way around.
In software, this often is similar to what happens. We buy software thinking it will save us time and money as it will improve the master/slave relationship to our daily lives. Instead, we become more enslaved in its processes.
An example. Today I went to my bank ANZ (which I am ditching - F*K you ANZ). I said, "I'd like a copy of my home mortgage statement to give to your competitor so I can leave your dumbasses - i.e. YOU ARE FIRED"
I watched the teller pound away at a keyboard for like 5mins before she arrived at a point of some kind that then needed her co-worker to give her instructions on generating a printable report.
I am sitting there thinking the following things:
- Why are you typing so much?
- Why can't I do this online myself? You give me access to every other account functionality yet why not this?
- Why am I giving you everything but a DNA sample to authenticate I am who I say I am still to this day?
My point here is that aside from a crappy online service from ANZ Bank, the teller herself should have a simple input control that has a button next to it. Inside that input control, she types, "Print <AccountNumberXYZ> Mortgage Statement as of Today"
The input box then does the following:
- Looks up my account number and verifies it still is active.
- Takes the verb Print to mean "fetch" and the words Mortgage Statement as being what should be fetched whilst the word "Today" meaning as in Now(). Then spit out a piece of paper with that information. In otherwords “PrintMortgageStatementWorkflow(custId, date);”
I think I make my point(s) in saying why are we jumping through hurdles to make software do the work when it feels like we are a separate background thread in the software's world.