What do you do when you see a ground-breaking product sell hundreds of millions of units and enjoy sky-high user satisfaction? You watch sales steadily drop after that initial wave, not because the fad is over but because the product is so good that people don’t need to replace it. It just keeps working great for years.

Welcome to Apple’s iPad dilemma.

Five years ago or so, the iPad’s sales were so strong that some predicted iPads would displace computers as the primary computing device for many people. That hasn’t happened, but it’s true that an iPad can be your primary computer for many activities, at least for hours at a time. I for one haven’t brought a laptop with me on business trips for years now, just my iPad.

But I still need my Mac or Windows PC; a full-on computer is what I use at my main workplaces, for the reasons we all know:

models. Styli are awkward for many screen manipulations outside drawing and checklists, which is why they have never taken off for general-purpose computing, even though they’ve been available for Windows and MacOS for decades. Apple really should support Bluetooth mice. How hard can that be? (Android does, after all.)

The more difficult issue is supporting larger screens. The iPad has long supported external displays, both through HDMI and VGA cables and through . So the issue isn’t the video connection per se.

The display challenge is having iOS adjust to a large screen. That means supporting multiple overlapping windows like Windows and MacOS have long done. It also means ensuring applications can work at any window size, again like Windows and MacOS. I have no doubt Apple can make this work—after all, Samsung has done it with the  for its Android-based Galaxy S8 smartphone. And I have no doubt that developers could adapt their apps to support such flexible large-screen windowing; because Apple’s Xcode development environment is used for creating both MacOS and iOS apps, it could even help them get there quickly.

) like .

iPads are widely used in several niches, such as by repair technicians, roaming sales clerks, and delivery staff. But they’re used largely as an adjunct device by information workers, something to bring to meetings and conferences by users secure in the knowledge their “real” computer is back at their desk or hotel room. The changes I describe could change that usage distribution broadly.

Of course, such an adaptive iPad would threaten MacBook sales, because the MacBook has evolved into a lightweight, power-sipping device that has matched two of the iPad’s initially revolutionary advantages over laptops: extreme portability and all-day battery usage. With the MacBook and Windows laptops having duplicated the iPad’s initially unique advantages, the iPad has struggled to justify its existence relative to a laptop for many users.

But , so it would be better for Apple to cannibalize MacBook sales through increased iPad sales than to let Windows tabtops and perhaps Dex-style Android devices do it instead. Apple made partial attempts to broaden the iPad’s utility with and then with the iPad Pro, a clear ripoff of the Microsoft Surface Pro. But Apple held off on going all the way.

The iPad is a great device. Apple just needs to make fit better into more people’s computing portfolios.