OS X Mavericks, Performance Anxiety & Phoning Home

| John Martellaro's Blog

There are several ways an OS can telegraph the status of its health and well-being. However, how that's communicated to the user, the developer and Apple are all very different things. Providing information to the Mac user that they can really use is a better approach. No intelligent agent required.

I've been thinking lately about how a modern OS, like OS X Mavericks, should communicate its status and to whom. The reason is that modern end users aren't really in a position to take technical steps when a UNIX operating system starts to act up.

IA: Sir, your Safari prefs are corrupted. This may cause a problem.
Answer: "Get lost!"

The User Dilemma

In the old days, UNIX IT admins would look at system logs (and some of us still do!) in order to identify a problem with the OS. Perhaps there is some conflict, a kernel extension is misbehaving or a preference is corrupted.

The problem is, if there were an animated, intelligent agent (IA) who could surface that problem to the user, there would be great anxiety about how to proceed. The response might be, "Okay, I don't even understand the warning. What do I do now?" Quizzing the agent further would likely take the discussion down a very bad rabbit hole. Like the infamous Clippy, the agent would risk becoming a running joke.


Developers have Xcode tools that help them optimize their code. They can see which subroutines are poorly written and perhaps chewing up too much CPU time. But, like Apple, developers know that there is not a lot they need or want to communicate to the user from under the hood. Customers just want their app to do the job at hand. So the communication to the developer, during coding, can be very geeky indeed but the customer never sees it.

A Better Approach

Awhile back, I suggested that future OSes have more self awareness of issues so that users are alerted to problems. Perhaps an intelligent agent. That's pretty geeky and also still fairly demanding of system resources. I've changed my mind.

The issue with that, I think these days, is that like apps, customers just expect their OS to work. If there's something wonky going on deep inside, say, due to bugs. the right way to deal with it is for the OS to communicate that back to Apple via the diagnostic settings. See, for example, "Why Apple Drops Features & How to Deal With it." Then Apple can fix the issue for all users, not just that particlar Mac.

That way, Apple can prioritize and fix issues that drive users crazy, users can get on with their work, and they won't have an OS that pesters them when it gets a headache and they have few good ideas about what action to take.

On the other hand, there are things the customer is all too willing to do. Such as evaluate the performance of an app, move an app into the background and chose what kind of workload, in general, they can place on the machine. For example, knowing which apps are using up the most battery power can help with a critical, on-the-fly decision while on a flight and no battery backup is available.

There's an old adage in local news: "News you can use." That's how OS X always has been and continues to be, and I think it will always be a good thing for Apple customers going forward. The only intelligent agent I want on my Mac is James Bond.


Intelligent Agent concept via Shutterstock.

Popular TMO Stories




As rational as is the idea you propose above, and as much as I concur with its practicality, methinks I descry a cadre of users who, discovering yet again that their software is communicating with Cupertino or their developer collaborators, will hoist the flag of revolt, draw their swords against the darkening shadows of conspiracy, denounce with vehemence and vitriol these silent communications, impugn the motives of said developers, and hold this as yet further evidence of Apple’s arrogance in removing yet further control of the system from the end-user - infantilisation at its most malevolent, they will charge.

And yet, should Apple and the developer community leave things as they are, and require end users to take steps to communicate either through web interface or in-app; many of these same souls will criticise Apple and all for failing to innovate more robust and fool-proof methods for communicating feature weaknesses to responsible parties without troubling the end user with such menial tasks.

Perhaps I’m mistaken, but I see a trend of increasing polarisation in nearly every sphere, in which its most curious feature is not so much the division of people over ideas and strategies, but the division of thought and mind within the same person depending on the initial choices with which they are presented.

Still, I welcome seamless and automated communication of my software directly with those whose role is to make it better.

Log in to comment (TMO, Twitter or Facebook) or Register for a TMO account