A few minor thoughts on Apple's new M1 processor

Apple has released their new M1 processor and their new Mac and Macbook computer which use the new processor. Apple is quite proud of this achievement (and they have every right to be) and users are impressed with the performance of the M1.

Yet I have some concerns about the new processors.

First, with the new processors and the new Macs, Apple has consistently emphasized performance, but little beyond that. Their demos have featured games and video editing, to show off the performance. The advertising is far from the presentations of yore which listed features and boasted MIPS ratings. (Perhaps video games are the new way to boast processor performance.) Apple's announcement was designed for consumers, not corporate buyers and not technologists and especially not developers. Apple has made their computers faster (and more efficient, so longer battery life) but not necessarily more capable.

Second, Apple's change to proprietary ARM processors (and let's be real, the design is proprietary) fractures the hardware market. Apple equipment is now different from "regular" personal computers. This is not new; Apple used PowerPC processors prior to using Intel processors in 2006. But the new Macs are different enough (and designed with enough security) that other operating systems cannot (and probably will never) be installed on them. Linus Torvalds has commented on the new Macs, and indicated that without Apple's support, Linux will not run on the M1 Macs.

Third, Apple made no announcement of cloud computing. This is not a surprise; Apple has not offered cloud computing and I see nothing to change that in the near future. It is telling that Amazon, not Apple, has announced new offerings of Apple M1 Mac minis in the cloud. Apple seems content to limit their use of cloud computing to iCloud storage, online documents, and Siri. I'm not expecting Apple to offer cloud computing for some time.

Moving beyond Apple, what will other manufacturers (Dell, Lenovo, etc) do? Will they switch to ARM processors?

I expect that they will not -- at least not on their own. Their situation is different. Apple has control over hardware and software. Apple designs the computer (and now, the processor). Apple designs and writes the operating system. Apple builds many of the applications for the Mac. This vertical integration lets Apple switch to a proprietary processor.

Manufacturers for "regular" personal computers rely on Windows, which is designed and written by Microsoft. Microsoft sells some applications, but encourages a large ecosystem of applications. A corresponding shift in the "regular" PC market requires the cooperation Microsoft, hardware manufacturers, and application developers. Such a change is possible, but I think Microsoft has to lead the effort.

Microsoft does provide a hardware specification for running Windows, and they could issue a specification that uses an ARM processor. Such a specification would exist along side the current Intel-based specification, expanding the space for Windows PCs. Microsoft could even design and build their own hardware, much as they did with Surface tablets and laptops.

I expect that Microsoft will support ARM processors (although not Apple's M1 processors) and offer cloud compute services for those processors. They may design a custom ARM processor, as Apple has done, or use a "standard" ARM processor design.

Getting back to Apple, I can see that the new M1 processor gives Apple more control over its market. The "expansion" market of running non-Apple operating systems is, as I see it, terminated with the M1 Macs. Boot Camp is out, Parallels works (but not with virtual machines), and non-Apple operating systems cannot be installed. (Technically that's not quite true. A clever engineer did get the ARM version of Windows to run on an M1 Mac, but only by using QEMU as an intermediary. It wasn't native ARM Windows on the M1 Mac.)

More control may sound good, but it does alienate some folks.

The Linux crowd will move to other hardware. Linux is a small presence in the desktop market, and Apple may not care or notice even. But Linux is a large presence in the developer market, and a decline in developers using Macs may affect Apple.

The Windows crowd has plenty of options for hardware, with more as Microsoft expands Windows to ARM processors. Windows laptops are just as good as Macbooks, and better in some ways, with options for non-glare displays and finger-friendly keyboards. My experience has shown that PC hardware tends to last longer than Apple hardware (especially equipment built after 2010).

I think, in the long run, Apple's move will fragment the market, aligning Apple hardware with Apple operating systems and Apple applications. It will push developers away from Apple hardware, and therefore Apple operating systems and the entire Apple ecosystem. Apple won't die from this, but it will give up some market share. The more noticeable effect will be the separation of processing on Apple and non-Apple platforms. Apple will be in their own world, their own garden of hardware, operating systems, and applications, and they will gradually drift away from mainstream computing.

Locked out of the walled gardens

A dark side of the walled gardens of technology is becoming visible.

The old paradigm was one of home ownership. You purchased your equipment (a PC) and then you installed whatever software you wanted. You decided. There were constraints, of course. PCs typically ran Windows (or before that, DOS) and the software had to run under the operating system. PCs had various configurations and the software had to fit (a floppy-only PC could not run large software that required a hard drive, for example).

Once installed, you had to see to the care and feeding of the software. You had to ensure that updates were applied. You had to ensure that data you exchanged with other users was compatible with their systems.

The benefit of this paradigm is the open market and the freedoms that come with it. Vendors are free to enter the market and offer their wares. You are free to choose among those products. You could pick from a number of word processors, spreadsheets, databases, compilers and IDEs, project managers, and other product categories.

The walled gardens of iOS and Android (and soon Windows and MacOS X) provide a different paradigm. If the old paradigm was one of home ownership, the new paradigm is one of renting an apartment. You still have a place for your stuff, yet a lot of the tedious chores of ownership have been removed.

With Apple's walled garden of iOS (and the gatekeeper iTunes), updates are automatic, and software is guaranteed to be compatible. The same holds for Google's Android garden and its gatekeeper 'Play Store'. They guard against 'unfriendly' software.

But the price of living inside the walled garden is that one loses the open market. Only selected vendors may enter the market, and those that do may offer a limited selection of products. Apple and Google enforce requirements for products in their walled gardens, through their registration and gatekeepers. Apple forbids a number of products in iOS and limits others. Web browsers, for example, must use Apple's WebKit engine and not install their own; Apple also forbids programming languages or scripting languages.

We're now seeing the Flash technology being pushed out of the walled gardens. Apple has prohibited it from the beginning. Google has deprecated it on YouTube. How long before Microsoft kicks it out of its garden?

The expulsion of Flash may foreshadow other exclusions of technology. Apple could, at any time, remove Microsoft's apps from iTunes. Google could remove Adobe apps from the Play Store. Microsoft could kick out Oracle apps from the (soon to be revived, I think) Microsoft App Store.

The ability to remove apps from the garden is something that the enterprise folks will want to think about. Building a business on a walled garden has the risk of existing at the whim of the gardener.

Centralizing or decentralizing? What's happening now?

One might be confused with the direction of today's PC technologies. Are they becoming more distributed or are they becoming more centralized?

There are two large-scale changes in today's technologies. Depending on which you look at, you will see the trend for distributed control or the trend for centralized control. (Kind of like those optical illusions of stairs that go up, or down, depending on your visual perception.)

The distribution of PC applications are becoming centralized: Apple's business model is centralized. Applications for iPhones, iPods, and iPads are distributed through iTunes and through iTunes only. This is very different from the approach used by Microsoft for PC-DOS and Windows applications, in which anyone could write and distribute an application, and anyone could purchase and install an application (provided they had administrator privileges) without any involvement or supervision from Microsoft. Apple has complete control over iTunes and one can distribute an iPad/iPhone/iPod app only with Apple's permission.

Apple has declared intentions to move the Mac world from the "open distributor" model to an "Apple centric" model with the "App Store". Indeed, applications that use iCloud must be distributed within the App Store. Applications distributed via the "open distributor" model cannot use iCloud services.

Microsoft is considering to closed distribution model with Windows 8 and the Metro environment.

The selection and ownership of PCs are becoming decentralized: The "bring your own device" fad (for now, let us call it a fad) shifts the ownership of PCs from employers to employees. The previous model saw employers specifying, providing, and provisioning PCs for workers. Often a company would have a standard configuration of hardware and software, issued to all employees. A standard configuration reduced support costs, since there was one (or a limited number) of hardware and software combinations.

With the "bring your own device" fad, employees pick the device, employees own the device, and employees provision and maintain the device. One person may pick a Windows laptop PC, another may pick a MacBook, and a third may pick a Linux tablet. This is clearly a decentralization of decisions -- although the employer retains decisions for the development of company-specific applications. An employer may develop custom software for their employees and build it for one or a limited number of platforms. (Such as custom software for insurance adjusters that runs only on iPads. You can be sure that insurance adjusters at that company will select iPads.)

The two shifts are symptoms of larger changes: Software is becoming a commodity, with the important packages running on multiple platforms (or equivalents running on different platforms). Second, power is shifting from PC customers (large user corporations) to PC platform manufacturers (Apple, Google, Microsoft).

Software is a commodity, and the different packages offer no compelling advantages. For word processors, Microsoft Office is just as good as Libre Office. And Libre Office is just as good as Microsoft Office. In the past, Microsoft Office did offer compelling advantages: it ran on Windows, it ran efficiently and reliably, and it used proprietary formats that demanded that new uses have the same software. Those advantages have disappeared, for various reasons.

Software is a commodity, and current products are "good enough". There is little to be gained by adding new features to a word processor, to a spreadsheet, to an e-mail/calendar application. I may sound a little like the "we don't need a patent office because we have invented everything" argument, but bear with me.

The core packages used to run offices (word processors, spreadsheets, e-mail, calendars, presentation, etc.) are good enough, the data is interchangeable (or convertable), and the user interfaces are easy enough to understand.

If we need additional functionality in an office, a company will get it (either by building it or buying it). But they will do so with extra software, not through extensions to the core packages. (The one possible exception might be spreadsheet macros.) The core office software are commodities.

Apple knows this. It spends no effort building its own version of office tools. I suspect that Microsoft understands this too, and is preparing for the day when lots of customers move away from Microsoft Office.

Apple and Microsoft are building new mechanisms to extract value from their customers: walled gardens in which they are the gatekeepers. The application software will be less important and the walls around the garden will be more important. Apple uses its iTunes as a tollbooth, extracting a percentage of every sale. I expect Microsoft to do the same.

What this means for "the rest of us" (individuals, user companies, developers, etc.) remains to be seen.

Why the Microsoft tablet for Windows 8 is significant

This past week Microsoft  announced a tablet for Windows 8.

Microsoft has a checkered history of hardware. Their successful devices include the Kinect, the XBOX, the "melted" keyboard, and the Microsoft Mouse. Failures include the Zune and the Kin, and possibly the Surface. Many people want to know: will the tablet be a success or a failure?

This is the wrong question.

Microsoft had to release a tablet. The market is changing, and the rules of the new market dictate that vendors provide hardware and software.

Microsoft must succeed with the tablet. Perhaps not this particular tablet, but they must succeed with *a* tablet, and probably several tablets.

The old PC market had a model that separated hardware, operating systems, and applications. Various vendors built PCs, Microsoft supplied the operating system, and multiple (non-hardware) vendors supplied application programs. In the early days, one purchased an IBM PC and PC-DOS (supplied to IBM by Microsoft), and Lotus 1-2-3 and WordPerfect. Once manufacturers figured out ways of (legally) building PC clones, one could buy a PC made by IBM, or Compaq, or Dell, or a host of other companies, but the operating system was supplied by Microsoft.

People who are knowledgeable of the history of computing will recognize that the separation of hardware and software originates in an anti-trust lawsuit against IBM, for mainframe software of all things. Yet that decision influenced the marketing arrangements for the original IBM PC, and those arrangements (software separate from hardware) influenced the entire PC market. (To be fair, the pre-IBM PC market had similar arrangements: Radio Shack TRS-80 computers, Apple II computers, and others let you add software -- any compatible software -- to the computer.)

That model endures with today's desktop and laptop PCs. We buy the PC, an operating system (usually Windows) is included, and we add our separately-acquired software. That software may come from any source, be it another company or our internal development shop. We are responsible for configuration and upkeep, and for problems due to incompatible software.

Apple, with the iPhone and iPad, uses a different model: they supply the hardware and the operating system, and while other vendors supply the applications, Apple limits the freedom of users to select application providers. With iTunes, only those apps approved by Apple can get onto an iPhone or iPad. Users cannot select any application, or any provider, or even have them custom-written. They must go through iTunes (and therefore Apple).

The phrase for this arrangement is "walled garden". The environment is a pretty place, but one cannot leave easily. The vendor has erected a wall around the garden, and occupants must remain within the constructed garden.

Walled gardens, especially in tech, have a downside. When leaving one garden for another, one often loses content. You can buy a dozen books for your Kindle. Buy a replacement Kindle and your books are available. Buy instead a Nook, and your books are not available. Barnes and Noble knows nothing about your purchases with Amazon.com, and Amazon.com has no incentive to make it easy (or even possible) to transfer your purchases to another device. While we can leave the garden, we do so only by  leaving things behind.

This walled garden model is used by game consoles. Games are made specific to consoles; the XBOX version of "Diablo 3" will run only on XBOX systems, not on Playstations. If you purchased an XBOX and lots of games, and then you decide to switch to the Playstation console, you don't get all the games available in their Playstation form.

We are entering an age of walled gardens. Apple has their iPhone/iPad garden, Amazon.com has theirs, Barnes and Noble is building theirs. With the introduction of the Microsoft tablet, we can see that Microsoft is building theirs. Google has built some infrastructure with Google Docs and the Chromebook laptop/browser.

This new age of walled gardens, of separate kingdoms, requires developers, users, and companies to make choices.

Developers must select the platforms to support. They can focus their efforts onto a limited number of platforms, or they can develop for all of them. But development for each platform requires tools, skills, and time. A large company can invest in multi-platform efforts; a small company with limited resources must choose a subset, possibly forgoing revenue from the omitted market segment.

Users must select their platforms carefully. The cost of changing is high, much higher than changing from a Dell PC to an Asus PC, or changing from a Ford to a Chevy.

Companies must select their platform. They have done so in the past, usually picking Microsoft, the safe choice. (How many times have you heard the phrase "we're a Microsoft shop"?) But now the choice is riskier; there is no safe choice, no one dominant provider (and no one provider that appears ready to become dominant). A company's success will depend not only on its talent and ability to execute business plans, but also upon the success of its platform. A company may do well in its market, yet fail when its technology provider fails.

These are difficult decisions, and they must be made. One cannot defer the selection of platforms; others in the world are moving to new platforms.

To wait is to be left behind.