In real life, there are plenty of doors for which you will never find the key lying around. More importantly there, are millions (billions?) of doors that are of no interest to you, ever. In a video game, it would be very difficult to set up a series of long term societal detriments for going around trying to open every door, or to easily express to the player why the character they're playing has no interest in one door vs. another, or why what's behind most doors is not of interest to the gameplay or the plot of the game. But it'd also be extremely strange to walk down a city street environment and have there be no doors into any of the surrounding buildings. So we put up false doors as window dressing so the environment looks familiar, but then we build a visual metaphor that lets players see at a glance which doors are unimportant so they don't bother to try them. This can be by leaving them as a flat texture instead of modeled, making openable doors a different color or have specific lighting or highlights, making openable doors have handles and unopenable ones not have handles, or as the article suggests, putting rubble or something (depending on the context of the game) in front of unopenable doors. You can even make unopenable doors make a specific sound effect when approached, such as the sound of a handle jiggling on a locked door, or the sound of the character specifically saying "It won't open," etc. (although only communicating it once the door is approached can be tedious for the player).
You're exactly right. Game developers can either spend a nightmarish amount of time trying to solve the locked door "problem", or they could make it a static prop that plays a "jiggle the doorknob, but it's locked" sound effect and then put that saved time and effort into the actual gameplay and mechanics. People sometimes forget that game developers have to pick and choose their battles, and any time spent on non-gameplay fluff should never detract from rock-solid core game mechanics.