A Question of Intent

Exploring the fractured relationship between intent and results in the software world

I've been giving a lot of thought recently to the way software designers, when creating a human-facing application, try to discern the human's intent.

There is a slippery slope when a software application goes from requiring explicit requests to assuming intentions based on actions or conditions.  The results can produce consequences that the human neither comprehended nor desired.

Everything Must Grow!
​
I will start with a simple, and personal, example.  Microsoft Windows, at some point along the way, added a "feature" which automatically resized windows based on certain dragging motions.  The one that drove me crazy was an automatic expansion to full screen mode whenever a window was dragged into the upper left corner area.

Note the threshold for this action is not absolute.  There is no marked target area to visually indicate where the logic will kick in.  As a result, many times when I would attempt to arrange a couple of windows to view side-by-side, the whole action would be stymied by one window taking over the screen.

There is a standard button on every window that effects this same behavior.  Any Windows user knows how to expand any given window to full screen with the click of a mouse button and/or a keyboard shortcut.  While a focused click may require a bit more concentration than an imprecise drag, the time requirements of each motion are pretty similar.

However, some development team somewhere, decided to add this interpretation of intent, in effect deciding that nobody would want to move a window into this area and keep it at its current size.  If you don't want to go full-screen, don't move a window into that area, wherever that area exactly is on your current resolution. 

The cost, in this case, was limited to personal frustration, until I hit my own threshold and researched how to disable it.

Social Distortion

The consequences increase when it comes to social media platforms.  At their core, most common social apps boil down to content creation and content distribution.  Each subscriber is both an author and a consumer of content.

My challenge to social media subscribers is this:  are you aware of, let alone in control of, all the content you create?  Some examples:

• You entered your birth date as a condition of subscribing (ostensibly for age eligibility checks).  Did you intend to have your birthday broadcast to your network every year?
• You clicked "like" on a friend's posting of some professional achievement.  Did you intend to create that entry in the content stream of your whole business network?
• You entered a date of employment to fill out your work history accurately.  Did you intend to alert your professional network annually on your "work anniversary?"
• You solved a math problem, presumably to prove you're a genius.  How many of your colleagues, friends and family did you intend to alert to this fact?

​On the consuming side of the content feed, you see the accumulated content creation events of everyone you know or to whom you subscribe,  Regardless of the creator's intent, the item is yours to consume.  Presently, the ability to filter or categorize content on most platforms is minimal -- you can filter someone entirely, or filter a few broad event types.  Beyond that, your feed is a flood of content, a significant percentage of which probably wouldn't be visible to you in the days before social media.

The platforms are motivated to get you to click something -- anything.  They are less motivated to fairly represent you or your intentions.  Asking you to explicitly declare your intent on content creation would slow you down, and most likely reduce your output.  Subsequently, less content created means fewer click opportunities for your connections and followers. 

It is really up to each individual to consider how much content they want to generate, both actively by explicit events, and passively by understanding what any given application will do with their data.

(Smart) Glass Houses

I am not casting aspersions on the authors of the specific software applications described here.  I am merely illustrating how easy it is for software to misrepresent a human's intentions, and create unwanted results.  The more shortcuts we add, the more streamlined we make something, the more assumptions get built into the software, and that's where disconnects happen.

When it comes to wearable technology, we have many more data points to consider, including computer vision, audio recognition, motion detection, even fatigue and vital signs.  Theoretically, we can make some very informed decisions on the wearer's intent as they navigate the physical world and a software application simultaneously.  However, as illustrated here, the challenge is to balance streamlined interaction with certainty of intent.  Stay tuned...