Lion Provides First Steps Toward Power Savings

In many ways, the trade press seems to have met the release of Mac OS X 10.7 Lion with a scratch of its collective head.

On one hand, after the long period of under-the-hood transition from the PowerPC architecture to Intel-based Macs that Apple undertook with 10.5 Leopard and 10.6 Snow Leopard, it’s good to see Apple focused, once again, on features that have a more tangible impact on the way we users interact with the operating system.

On the other, there is no doubt that Lion asks a lot of its users — sometimes with little in the way of explanation. Many have commented negatively on features like the reversal of the scroll direction, user interface tweaks, Mission Control’s changes from Spaces, and so on (see “Subtle Irritations in Lion,” 17 August 2011). And like Matt Neuburg’s experience in “Lion Is a Quitter” (5 August 2011), I’ve found myself wondering where my apps have disappeared to, and my first hour of work with the latest release of Pages was punctuated by frantic attempts to figure out what had happened to the File menu’s “Save As…”
item.

Lion, in other words, is asking us to change habits that we have formed over twenty-five years of graphical user interface usage — and, true to form, Apple hasn’t told us why we would ever want to do this.

This is particularly hard on power users (who, I would guess, comprise the majority of the trade press), who have learned to maximize productivity by anticipating Mac OS X’s behavior at every turn. Decades of consistency have created a muscle memory that Lion confounds at every turn, from scrolling to saving and switching between apps.

The history of computing doesn’t lack for oddities and bad decisions — some of which have come from Apple itself. In the past few years, however, the company has had a remarkable streak of successes that reveals a methodical approach to innovation. As a result, it seems overly simplistic to simply say “Apple has biffed it” and chalk Lion’s strangeness up to incompetence — or, perhaps, to an obscure form of brain fever caused by having nearly $80 billion in cash. (No matter how realistic we try to be, we can’t shake the feeling that Apple has, somewhere in the depths of 1 Infinite Loop, a Scrooge McDuck-like money pit for executive wallowing.)

It seems to me that a more likely explanation for Apple’s decisions with Lion requires additional analysis, with particular attention paid to where Apple is today and to where it might be charting its own future. The answer to Lion’s quirks, I believe, lies in a disconnect between Macs and the company’s mobile strategy.

There is much to like in Apple’s iOS devices, but one of their killer features is their combination of portability and longevity. The iPad may not replace a laptop, but it provides an excellent balance between functionality and battery time; it is still the only general-purpose computer that can make any intercontinental flight between America and Europe on a single battery charge.

What’s interesting is that the technology behind this remarkable feat is not related primarily to the device’s hardware; there is little inside an iPad that wasn’t available to other manufacturers before its introduction, and yet none of them (including Apple) has ever been able to create a laptop with a comparable balance between features and power consumption.

The iPad’s real achievement is in its software; iOS is not just a beautifully responsive operating system that has revolutionized the way we interact with mobile apps, it’s also an environment whose unique constraints have made it possible for Apple to create devices with unrivalled battery life.

For instance, you may recall that, when the iPhone was first introduced, Apple was widely derided for the lack of “multitasking” capabilities, something that practically every other operating system — including Mac OS X — had been offering for years.

Even the first version of iOS was, of course, fully capable of running more than one app at the same time — and, in fact, did so with a subset of Apple’s own apps. Apple had just chosen to prevent users from keeping multiple apps active because allowing only one app to run at once allowed the iPhone to use less RAM and power, resulting in both lower manufacturing costs and higher battery performance.

Fast forward a few years, and the hardware had evolved to the point where Apple decided to allow multiple apps to run concurrently. Even then, however, the company did so in a highly controlled manner: each app would be allowed to perform one of only a handful of officially recognized actions while in the background. All other activities are stopped by iOS’s watchdog process as soon as the user switches away from the app.

You can see, therefore, that the real breakthrough that made the iPad, iPhone, and iPod touch such amazing traveling companions wasn’t a new chip or kind of battery — it was the software itself.

Now consider Apple’s other mobile platform: the laptop. According to the latest numbers out of Cupertino (see “Apple Reports Q3 2011 Record Financial Results ,” 19 July 2011), MacBooks now outsell the company’s desktop models nearly three to one, making them Apple’s most important non-iOS line of products.

It’s not surprising — as our lifestyles have become increasingly mobile, MacBooks have evolved considerably. Faster processors, better screens, and amazing new construction technologies have contributed to making them the best portable computers on the market.

One feature that hasn’t improved dramatically, however, is the battery life. Even the introduction of the unibody models, in which the battery is built into the laptop’s chassis, has given us only “up to 7 hours” of advertised longevity (and the 11-inch MacBook Air gets less than that, at “up to 5 hours”). Plus, while Apple surprised the press when the iPad’s advertised battery life turned out to be shorter than what practical tests revealed, the MacBook turns in highly variable results depending on the kind of testing performed.

Short of an unexpected breakthrough in battery technology, therefore, Apple has probably realized that they’ll have to rely on software to give MacBooks the same battery life that has helped make the iPad so popular — and this is where many of the new features of Lion come into play. Also, although improved battery life is something that every MacBook user would appreciate, reduced power consumption for desktop Macs would be welcome at least at a societal level, even if it wouldn’t make a big difference in any individual user’s power bill.

Many of the under-the-hood changes that have been introduced in Lion are meant to shift the control of apps from the user to the operating system. Thanks to features like Automatic Termination, Auto Save, and Resume, Lion is increasingly the arbiter of when and how an app gets to run.

We certainly are not there yet, but it’s possible to imagine a future in which all these technologies work together to provide users with notably improved battery life. For example, if Mac OS X detects that an application is simply idling, it could cause it to save all its data automatically, quietly terminate it, and then transparently bring everything back when requested. That works in iOS today, and could work in Mac OS X soon.

Similarly, apps that currently use background threads for everything from checking mail to keeping your Internet chats going could take advantage of push notifications to delegate much of their workload to the operating system, reducing both their CPU usage and power requirements.

Better yet, the entire world of computing could return to a time when we all ended our workdays by turning off our Macs. Combining Lion’s Resume feature with the speed of an SSD, the time difference between a cold boot and waking up from sleep could be so small as to be irrelevant. And although the power difference may not seem large, recent research shows that appliances in standby mode are responsible for hundreds of thousands of tons of carbon emissions every year.

As is often the case, each individual power-saving feature may have a small impact on a Mac’s overall power usage, but together they could result in much longer battery life — and in batteries that last longer before needing to be replaced. It’s a little like a hybrid car that turns its engine off while idling at a light and uses regenerative braking to charge its batteries; neither is huge in its own right, but the cumulative effects add up.

Obviously, not all the pieces of this puzzle are there today: most developers haven’t yet figured out how to integrate all of Lion’s new features; hardware changes — like switching to solid-state drives and possibly even more-efficient processors like Apple’s A5 — will have to be made across the entire MacBook line; and, judging from the early reactions to Lion, Apple itself still needs to figure out exactly how some of the lessons it has learned from iOS apply to a desktop environment.

For instance, although there are some who have seen improved battery life after upgrading to Lion, a large number of people are complaining in a lengthy Apple discussion thread that Lion has noticeably hurt battery life. There are some basic fixes, such as resetting the SMC and using Cody Krieger’s gfxCardStatus to make sure that MacBook Pros use the more-efficient integrated graphics when possible. But the hardware site AnandTech found that while Lion performed slightly better than Snow
Leopard in terms of battery life on an SSD-equipped early-2011 15-inch MacBook Pro, it was about 20 percent worse on an early-2008 15-inch MacBook Pro. Even some Lion features that could eventually result in less power usage in the future aren’t working properly now, so, for example, when Lion automatically terminates what it considers to be an unused application, removing its icon from the Dock and the app switcher, the app’s process continues to live on, as you can verify in Activity Monitor. Apple would seem to have this backwards from both the user interface and power savings standpoints — the process of an automatically terminated app should exit, but its representation in the Dock and app switcher should remain.

Nonetheless, even after Apple fixes the parts of Lion that are broken and continues to evolve the concept of software-based battery life improvements, we users will still need to adjust. My suspicion is that that’s Apple’s goal with Lion — to help Mac users become comfortable with the idea of documents that save themselves, apps that are running but not running, and an operating system that attempts to anticipate our usage patterns. We may not see the full effect of Apple’s power-saving strategy until later releases of Lion, or even until the next big cat. But it is coming, and while we long-time users may have some trouble understanding this new approach toward saving power, it also took some time for many people to become
comfortable with the Mac itself, with its paradigm-shifting graphical interface. Life is change, and the means may not always be the most comfortable way of achieving the ends.