The Windows 10 Upgrade and Alternatives

This post discusses the one-year free upgrade offer from Windows 7 or 8 to Windows 10, explores Microsoft’s reasons behind that offer, and addresses the question of how PC users might want to respond. Microsoft’s reasoning may be relevant for those who want to base their operating system choices on an informed sense of what the future may bring.


The situation appeared to be as follows: Microsoft offered Windows 10 as a free upgrade, and pushed users of Windows 7 and 8 to accept it, because Microsoft considered Windows 10 important for the company’s future: (1) Windows 10 was crucial in Microsoft’s bid to control the Internet of Things, with billions of devices running Windows 10 IoT; (2) Microsoft wanted to sell a very stable Windows 10 to businesses, and planned to use millions of Windows 7 and 8 users as guinea pigs who would be force-fed the Windows 10 operating system and its potentially flawed updates, so as to catch bugs before they reached corporate users; and (3) Microsoft designed Windows 10 to produce revenues by obtaining detailed personal information on users and their computing activities.

Decisions incorporated into Windows 8 and 10 suggested that Microsoft increasingly took desktop and laptop computing for granted. The company’s attempts to shoehorn a single operating system into many different devices (e.g., smartphones, tablets) made it more difficult for PC users to get work done. At the same time, new Intel CPUs were designed not to work with Windows 7. Microsoft’s high-pressure efforts to drive users out of Windows 7 and into Windows 10 backfired in some cases, as users considered alternatives that might yield a long-term solution based on a well-organized and trustworthy operating system. Linux, in particular, was able to run on those new Intel CPUs; was able to run many Windows programs or functional equivalents; and also offered the ability to run the entire Windows (7, 8, or 10) operating system inside a virtual machine — all without the prospect of Microsoft’s plans to sell its software on a more expensive subscription basis. The Conclusion offers a more detailed summary.



The Free Upgrade Offer
The Windows 10 Life Cycle
Windows As a Service
The Supported Lifetime of the Device
The Pace of Windows 10 Adoption
Negative Responses to Windows 10
Advantages of Windows 10
Computers: A Rumored Death Greatly Exaggerated
The Hard Sell
The Microsoft Mentality
The Pivot to the Cloud
Why Did Microsoft Need Windows 10?
Microsoft’s New Priorities
Where’s the Money?
What Could Happen Next
The End of the Free Upgrade


Short-Term Fixes
A Longer-Term Alternative: Windows 10 Enterprise LTSB
Non-Windows Alternatives
Windows Programs or Linux Counterparts
Running Windows Inside or Along With Linux




The Free Upgrade Offer

Microsoft said that, as of July 29, 2015, Windows 10 became “available as a free upgrade to people using Windows 7 and Windows 8.1.” The footnote accompanying that press release pointed to another page with more details. Perhaps the most important of those details: “To take advantage of this free offer, you must upgrade to Windows 10 within one year of availability.”

That raised several questions. First, what was going to happen when that year expired on July 29, 2016? A Microsoft Community post interpreted Microsoft executives as implying that users who qualified for the free upgrade to Windows 10 might be paying for updates after July 29, 2016. But that seemed unlikely. Microsoft Executive Vice President Terry Myerson said, “Once a Windows device is upgraded to Windows 10, we will continue to keep it current for the supported lifetime of the device – at no cost.”

The Windows 10 Life Cycle

That plan to “keep it current” evidently meant “a commitment to deliver free, ongoing security updates.” But not forever. A Windows lifecycle fact sheet indicated that mainstream support for the version of Windows 10 released in July 2015 would end on October 13, 2020, and extended support would end on October 14, 2025. A Microsoft Support Lifecycle webpage explained that mainstream support included product improvements and certain other kinds of support and assistance, while extended support was largely limited to security updates.

Those dates implied that there was going to be a Windows 11, and that it would be coming out sometime before that 2020 date. But that was incompatible with the statement, by Microsoft executive Jerry Nixon, that “Windows 10 is the last version of Windows.” The explanation could be simply that the next Microsoft operating system would be called something other than “Windows.” According to BBC, interpreting remarks by industry analyst Steve Kleynhans, “The company said it had yet to decide on what to call the operating system beyond Windows 10.” It was also possible that the name of this Post-Windows 10 operating system would just be something like Windows 10.1.

The Windows lifecycle fact sheet indicated that mainstream support for Windows 8 was scheduled to end on January 9, 2018. The July 2015 release of the next edition, Windows 10, came about 2.5 years before that termination date. Similar timing in the future would suggest that the Post-Windows operating system, whatever its name, would be released sometime around April 2018, 2.5 years before the October 2020 end of mainstream support for Windows 10. But Microsoft would probably begin talking about Post-Windows even earlier: the company officially announced the name of Windows 10 in September 2014, ten months before its release. Ten months before April 2018 would mean that one could begin hearing official Microsoft announcements about Post-Windows in early 2017.

Of course, Windows 8 was a flop. The company plainly felt a need to get past it sooner rather than later. The foregoing dates might be postponed if Windows 10 succeeded.

Not surprisingly, rumors of a future “Windows 9” commenced even before the October 2012 release of the tragic Windows 8. As soon as Windows 7 users got a look at it, many decided to stay where they were. Similarly, some would no doubt dislike Windows 10. One could expect Microsoft to try to persuade them to accept it anyway. They would be less inclined to do so if they believed that a Windows 11 would be coming along.

That appeared to explain Nixon’s statement: “[B]ecause Windows 10 is the last version of Windows, we’re all still working on Windows 10.” That sounded like a claim that you must accept Windows 10 because there will be no more Microsoft operating systems after Win10. But it appeared actually to be a statement that we want you to accept Windows 10 because we are not going to talk about Post-Windows yet.

Windows As a Service

Many articles referred to the concept of “Windows as a service” (WAAS). Microsoft did not appear to have presented this concept in great detail. A Microsoft blogger described it as “the new model of rolling out steady, predictable updates, features and functionality to the Windows 10 OS instead of monolithic version changes” with “continual, on-going OS innovation on a given device over time.” As Ars Technica put it, “New features and capabilities will no longer be held back until ‘the next major release of Windows'”; instead, they would be shared with users as soon as they were finalized.

The claims that there would be no more “monolithic version changes” or any “next major release of Windows” seemed to imply that Microsoft was giving up on a former source of funds. The Ars Technica article contended that this was not a significant concern — that sales of Windows upgrades “were never a major revenue source” for Microsoft. But there seemed to be more at stake than just the sale of upgrade licenses to Windows 10. In the past, the appearance of a new Microsoft operating system had also stimulated purchases of new computers, as people got excited about new capabilities and worried about being left behind. By contrast, according to the IDC market research firm in August and October 2015, PC shipments were down significantly due to “the availability of a free upgrade to Windows 10” and seemed likely not to grow significantly until 2019 “as saturation and ‘good enough computing’ sentiments spread.”

In other words, Microsoft would not only not be selling as many upgrade licenses; its free Windows 10 upgrade strategy appeared to be reducing its revenues from new-machine licenses as well. TheStreet said that, as of September 30, 2015, at Microsoft, “Overall sales of Windows operating systems declined by 17% from a year ago,” though some of that was no doubt due to a general slowdown in the industry.

Despite the belief that Windows 10 meant “an end to the huge, bundled service packs of yore” (ComputerWeekly, 2015), it did not appear, in practice, that WAAS had truly changed Microsoft’s approach to operating system updates. At this writing, the company was planning a major new Windows 10 update known as Redstone. This update, rescheduled for early 2017, was not entirely consistent with what ZDNet described as Microsoft’s plan “to push two to three ‘major’ collections of new features to Windows 10 users annually.” The ZDNet article linked this Redstone update to “a new wave” of Microsoft device offerings. It seemed, in other words, that marketing and management considerations would continue to drive strategic deployment of major updates, as distinct from a constant dribble of incremental changes.

If WAAS permitted Microsoft to offer “major” updates to Windows 10, and to postpone them for as much as a year, then apparently WAAS would permit Microsoft to charge users for a Post-Windows operating system, whenever the company might next find a persuasive rationale for selling such an upgrade. But some felt that this would not happen — that Windows would remain free forevermore. ComputerworldUK (Finnegan, 2016) quoted Gartner analysts Kleynhans and Silver for the view that the market had changed permanently:

Microsoft is the only remaining vendor that directly tries to monetise a client device OS.

Users have become conditioned to seeing the OS as part of the device, and something that should just get updated for the life of that device. Microsoft’s approach of charging for upgrades has seemed out of step.

On the other hand, Microsoft enjoyed a dominant position within the PC operating system market (below), and as such would probably be able to find ways to sell versions of it to at least some customers. Indeed, ComputerworldUK observed that this new Windows 10 operating system would be free only for consumers and small businesses — not for the Enterprise version.

The Supported Lifetime of the Device

Myerson’s statement (above) was, “Once a Windows device is upgraded to Windows 10, we will continue to keep it current for the supported lifetime of the device – at no cost.”

In one sense, that statement appeared redundant: “the supported lifetime of the device” seemed to mean simply “the period of time during which Microsoft would continue to keep Windows 10 operational on that device.” That is, support = keeping it current. The sentence could apparently be rephrased as, “Windows 10 devices will receive free Windows 10 upgrades until Microsoft decides to stop supporting those devices.”

The question, then, was when Microsoft would decide to stop supporting a device. According to the Servicing Option Summary at the bottom of a Microsoft Technet webpage, WAAS meant that (aside from Enterprise edition) Windows 10 devices would be supported for a minimum of four months (for general users) or eight months (for business users).

Ordinarily, one would expect desktop computers to be supported for some years. I saw, for example, that the Windows 10 hardware requirements — such as a 1GHz CPU, 2GB RAM (for 64 bit), and DirectX 9 — were largely available on desktop machines more than ten years before the Windows 10 release date. But other developments, discussed below, posed doubts about this expectation. (Note: for computer owners considering hardware upgrades, it appeared that the license was tied specifically to the motherboard.)

The Pace of Windows 10 Adoption

According to Netmarketshare, as of October 2015, the desktop operating system market was substantially divided among Windows 7 (56%), Windows 8 (13%), Windows XP (12%), Windows 10 (8%), Mac OS X (8%), Vista, NT, and older versions of Windows (<2%), and Linux (<2%). By February 2016, the percentages had changed somewhat for the newer Windows versions: Windows 7 (52%), Windows 8 (13%), Windows 10 (12%), and Windows XP (11%). Those figures raised several thoughts:

  • The October 2015 and February 2016 figures suggested a very slow embrace of Windows 10.
  • Windows 8 figures remained nearly unchanged, suggesting that those who had gone that route mostly lacked the need and/or desire to plunge into yet another new Microsoft operating system.
  • The small rise in Windows 10 adoption between October and February came mostly from Windows 7 and secondarily from Windows XP.
  • Among computers running some version of Windows, 58% were still running Windows 7 as of February 2016, six months after the release of Windows 10.

StatCounter (2016) offered somewhat different figures for desktop operating systems. For July 2015 and March 2016, those figures were as follows: Windows 7: 55%/46%. Windows 10: 0%/16%. Windows 8: 21%/15%. Windows XP: 10%/8%. Mac OS X: 9%/10%. There were very slight declines in the tiny Vista and Linux percentages during that timeframe, and a rise of about 1% in Other operating systems. These and other data offered on that StatCounter webpage provoked several additional observations:

  • At the time of the most intense Windows 10 promotion, the introduction of Windows 10 appeared to be linked with small rises, not declines, in the preference for Mac and some other non-Microsoft operating systems.
  • As remarked by Forbes (Kelly, 2015), the rate of Windows 10 adoption slowed dramatically within a few months after the July 2015 release. Between November 2015 and March 2016, Windows 10 achieved only a 1.4% average monthly rise in its share of all desktop operating systems. In that same period, Windows 7 dropped by only 0.8% per month. At those rates, Windows 10 would not achieve the market share now held by Windows 7 until 2018, and Windows 7 would continue to enjoy popularity exceeding that now enjoyed by Windows 8 until 2020.
  • The rate of change had slowed even more in the past few months. Between January and March 2016, Windows 7 declined by only 0.3% per month, and Windows 10 climbed by only 1.0% per month. At those rates, it would be necessary to add a couple of years to each of the preceding projections. Basically, Windows 10 would not have vastly superseded Windows 7 by the scheduled end of Windows 10 mainstream support in 2020.

Microsoft reported that that number of devices running Windows 10 had risen to 200 million by the end of 2015. Note, however, that “devices” would include non-computing devices in the Internet of Things (below). Computerworld (Vaughan-Nichols, 2016) pointed to the federal government’s Digital Analytics Program (DAP), which reported that, in the 90-day period ending on March 9, 2016, the operating systems used by devices (of any kind) visiting government websites were as follows: Windows 7, 33%; Windows 10, 10%; Windows 8, 7%; Windows XP, 2%; iOS, 19%; Android, 18%; Macintosh, 8%; all others, 3%. These numbers prompted additional observations:

  • The DAP percentages might differ at certain times of year. At this writing, during the U.S. tax season ending April 15, DAP indicated that federal webpages related to income taxes were the overwhelming favorites. The devices that people used to prepare income tax returns might not be the devices they would use most frequently at other times of year.
  • Data from Netmarketshare indicate that the Android and iOS mobile device operating systems accounted for virtually the entire market for mobile and tablet operating systems. Those two operating systems accounted for 37% of government website visits in the past 90 days.
  • Among major non-mobile operating systems, the DAP data indicated that visits to government websites within the past 90 days were allocated as follows: Windows 7, 57%; Windows 10, 17%; Windows 8, 12%; OS X, 15%. As noted by Vaughan-Nichols, these numbers more reasonably relegate Windows XP to the dustbin. Otherwise, aside from a possible tax/wealth effect related especially to Apple operating systems, these seemed roughly similar to market share percentages cited above.

Computerworld (Keizer, 2016) stated that corporations were expressing unprecedented interest in the upgrade to Windows 10. But Computerworld also indicated that large-scale corporate deployments would probably not take place until 2017 or 2018 and would be driven by a desire to complete the transition to Windows 10 before the end of Microsoft’s support for Windows 7 in 2020. That did not sound like enthusiasm. Large companies would have to allow a fair amount of time for what Computerworld described as “piloting and early deployment,” given that Windows 10 seemed to have been rushed out without too much prior collaboration and consultation to insure that Microsoft was really meeting corporate needs. Hence, the timing suggested a corporate inclination to postpone the Windows 10 upgrade until all other timely options were exhausted.

As in previous upgrades, it appeared that plain-vanilla users would face the fewest hurdles in transitioning to the new operating system. That would presumably include organizations using legions of largely identical computers running standard software. For instance, the U.S. Department of Defense announced a decision to upgrade 4 million systems to Windows 10 by early 2017. But myriad users with more diverse hardware and software who faced no pressing need to upgrade might hesitate, lest the upgrade process leave them without access to needed tools.

Negative Responses to Windows 10

The disaster of Windows 8 had reminded Microsoft that an operating system upgrade could flop, in the sense that many would decline to upgrade, and many who did would be unhappy. There was a fair question as to whether Windows 10 would do any better.

On that question, let me start with my own experience. On January 1, 2011, I decided to stop using Ubuntu Linux. I had gotten sick of crashes and other problems in Windows XP, and had thus been using Ubuntu for several years. I found, indeed, that WinXP, running in a VMware virtual machine in Ubuntu, was actually more stable than WinXP running natively.

I switched back from Ubuntu to Windows in January 2011 because, by that point, the situation was clear: Windows 7 was a winner. And you know what? I think it still is.

According to what some call the alternate versions rule, alternate Windows versions rule: one is poor, the next one is good. This has arguably been the situation since the 1990s. I did not find Windows 3 superior to IBM’s OS/2 or, for that matter, to DOS, given the DOS-based GUI programs (e.g., WordPerfect) that began to appear late in the game. But after Windows 95, Win98 was a real improvement, Windows ME was useless, WinXP was another improvement, Vista wasn’t, and then there were Windows 7 and 8.

That rule suggested that Windows 10 would mark a retreat from the worst aspects of Windows 8, and would also provide greater stability and functionality than Windows 7. And perhaps it did, or someday would. But I was writing these words after spending days working through various aspects of an upgrade from Windows 7 to Windows 10. For me, that upgrade consisted of two parts: the basic installation, and then customization so that I could use Windows 10 as I had been using Windows 7. As described in other posts, the upgrade was a convoluted struggle that left me with several issues that I could not resolve, and I suspended the tweaking effort after investing an inordinate amount of time and still finding myself frustrated by operating system characteristics that seemed hostile to the things that a person like me might want to accomplish.

I was not alone in having such reactions. For example,

  • ExtremeTech (Cardinal, 2015) described an inordinately frustrating and time-consuming process of upgrading from Windows 7 to 10, referring to “horrible parts of Windows Update,” error messages that are “completely inscrutable,” and “the next insane thing Microsoft has done that will drive Windows 7 users crazy.”
  • TechTarget mentioned potential incompatibility with key hardware and software, and suggested holding off on the Windows 10 upgrade until Microsoft had caught and fixed more of its bugs. TechSpot (Walton, 2015) and MaximumPC (Campbell, 2015) also found numerous bugs.
  • The Free Software Foundation (FSF) “urge[d] everyone to reject Windows 10 and join us in the world of free software” because, among other things, in Windows 10, Microsoft “is reported to give the NSA special security tip-offs that it could use to crack into Windows computers,” as detailed in a BloombergBusiness article.
  • FSF also said that Microsoft’s privacy policy “asserts the privilege to sell almost any information it wants about users.” I had struggled with privacy settings during my own installation and tweaking efforts, and had seen indications that even those efforts could not entirely turn off the flow of personal data to Microsoft.
  • FossForce examined the Windows 10 policy and license, and concluded that it enabled Microsoft to force updates on users, including updates that users might not want — updates that might even break existing functionality. InfoWorld (Leonhard, 2015) noted that, by midyear 2015, there had already been about 40 problematic Windows updates. In a slideshow, Leonhard recalled the 20 worst Microsoft updates of all time, some of which made computers unbootable, disabled functionality in the week before annual tax returns were due, and resulted in lawsuits. Forbes (Kelly, 2015) reported that, even before the official release of Windows 10 in July 2015, its automatic updates had already begun to cause many PCs to stop working correctly.
  • Laptop reported efforts by Dell and HP representatives to discourage users from upgrading to Windows 10, though this was not consistent with those companies’ official positions. Executives explained that technicians might encourage a rollback to an earlier version of Windows in order to correct user problems in specific cases. ZDNet said the techs’ reaction wasn’t surprising: “PC OEMs [Original Equipment Manufacturers] already operate on razor-thin margins, and asking them to take on the job of supporting upgrades is unreasonable.”
  • Many users expressed dismay with Microsoft’s decision to remove Windows Media Center from Windows 10.
  • At the time of the Windows 10 rollout in July 2015, Computerworld carried an article titled “Windows 10 is for suckers.” Reasons included “a ton of bugs,” software and hardware incompatibility issues, poor design choices, security problems, and “the distinct impression that Windows 10 is being rushed out the door, before it’s fully baked.”
  • As examples of incomplete development, DigitalTrends (Smith, 2016) referred to “feature sprawl” (in e.g., the combination of Settings, Control Panel, the “hamburger” menu, Action Center, and Notification Center), and said, “There are still menus that look like they belong to another operating system, or early alpha software, and somehow snuck into the release.”
  • TrustedReviews (Chester, 2016) said, “Overall, we’re still unconvinced that the new styling is actually better than the old.”
  • Computerworld concluded with advice from Susan Bradley: “wait to install Win10 on your primary or production PCs … Let tech writers and equally insane Windows users install Win10 first and report what they find.”
  • After the much-anticipated first major upgrade in November 2015, InfoWorld (Leonhard, 2015) reported “only tiny steps” toward resolving “the many holes” in the July 2015 release, and concluded,

Thus far, there is simply no compelling reason to switch from Windows 7. . . . It was easy to give Windows 10 RTM a vote of confidence when it originally shipped on July 29, but now that we see how slowly the changes are coming, the enthusiasm is starting to wane. It’s getting harder to envision a future where Windows 10 is a platform for Universal apps across phones, tablets, and desktops — and even harder to imagine a future where app developers give two hoots about Universal apps.

At some point, we’ll have to bite the bullet and switch simply because Windows 10 is new and Windows 7 is old — not because Windows 10 is better.

  • Brandwatch Blog examined the trend in user reactions during the first week or so after the Windows 10 launch. The initial euphoria quickly declined. By August 5, 2015, people who loved Windows 10 still outnumbered those who hated it by about 3 to 1 — not horrible, but a great decline from the initial 10:1 ratio. The features that some people liked most about Windows 10 were, for the most part, the very features that (according to Brandwatch) others hated most about it: the Start Menu, the Edge browser, Cortana, Xbox integration, the Action Center, and the Taskbar.
  • Danylko described his Win10 installation experience. Included among his remarks: “It got to the point where if I used my computer for more than 15 minutes, it would ‘blue-screen’ on me and reboot. His advice was to “Set aside a good chunk of time to make sure nothing happens with your upgrade.” But he also concluded that a clean install worked much better than an upgrade install, and that those doing a plain-vanilla installation might have few problems.
  • MakeUseOf presented a list of things that they wished Windows 10 had done right. These included leaving it to users to decide whether to install updates; the confusing mix of Control Panel and Settings; a user interface mishmash; privacy invasion; Cortana’s limitation to just a few countries; and deficient parental controls.
  • Breitbart listed reasons why people with conservative political or economic views might find Windows 10 unpalatable. These included Microsoft’s anticompetitive efforts to force OEMs to offer only Windows 10 on the machines they were selling, or to make their hardware compatible only with Win10, when they and their customers would love to have more Win7 machines; invasive data collection; default peer-to-peer impositions on user bandwidth; letting social media “friends” sign into one’s network by default; unreliability; and change for the sake of change, making it difficult for people to do the things they had previously been able to do, as they went about living their lives and striving to fulfill their potential.

There appeared to be quite a few other sources expressing frustration and dislike toward Windows 10.

Advantages of Windows 10

Obviously, a lot of people were making the transition to Windows 10, and many were happy with it. A search led to the following details regarding various claimed and real advantages of Windows 10:

  • Forbes (Kelly, 2015) said that Windows 10 had the advantage of being free. But this would matter only if Windows 10 offered something that users would otherwise be willing to pay for, or would wish they could afford.
  • Forbes also said that Windows 10 had the advantage of being supported by Microsoft for a longer period of time than Windows 7 or 8. Of course, this was an artificial advantage, created by Microsoft for purposes of persuading people to shift to Windows 10. Microsoft could instead have extended the support period on Windows 7.
  • Forbes stated that Windows 10 would have the ability to run on all future Microsoft devices. That became an advantage only when Microsoft decided to prevent Windows 7 from running on future devices. If Microsoft had been motivated by concern for device compatibility, it would have focused on improving Windows 7, with its larger and more established set of device drivers and utilities.
  • According to InfoWorld (Leonhard, 2015), “Windows 10 is the way of the future.” Specifically, Microsoft’s “universal apps” strategy, built into Windows 10, would make it easier for developers to write code that would work on a variety of devices and even on other operating systems (notably Android, iOS, and Mac OS). This could help Microsoft become a participant in the mobile market dominated by Google and Apple. But for people working primarily in a desktop environment, this strategy might prove useful only at the future time (if any) when they would encounter universal apps capable of doing things that they could not do with regular Windows programs that might exist now or be developed in the future.
  • Forbes raised the question of whether an operating system designed to run on smartphones would continue to be optimized for desktops and laptops, and vice versa. As ExtremeTech noted, the reality of people using a variety of devices for diverse purposes would continue to oblige manufacturers and software developers to tailor their offerings to specific needs and hardware. It seemed, then, that the Windows 10 attempt to be all things to all people could entail complications and performance issues that might be better addressed by using separate operating systems targeted toward very different kinds of uses and interfaces — with, perhaps, a somewhat standardized user interface.
  • Forbes cited DirectX 12 as an important advantage of Windows 10 over previous versions, especially for gamers. PCWorld (Chacos, 2015) observed that Microsoft had decided not to make DirectX 12 available on prior versions of Windows. This seemed to be another artificial advantage, created by Microsoft to compel an upgrade to Windows 10 but not necessitated by any core limitations in the Windows 7 operating system. PCWorld reminded users that DirectX 12 would show a significant performance difference only in programs written specifically for DirectX 12.
  • DirectX aside, there did not appear to be significant performance advantages in Windows 10. TrustedReviews (Chester, 2016) spoke of faster boot-up time and a generally “nippier” feel, but a TechSpot review (Walton, 2015) found “near-identical results . . . on most tests, from application to encoding, to storage and gaming” when comparing Windows 7, 8, and 10; likewise MaximumPC (Campbell, 2015).
  • Forbes said that Windows 10 had the advantage of offering the Cortana voice response feature. But, of course, voice command technology was already offered, by Microsoft and by third parties, in Windows 7, and the same was true for searching within the names and contents of files on one’s computer (see e.g., Everything; Copernic) — and these tools did not entail sharing of one’s personal data with Microsoft, advertisers, and law enforcement. Moreover, Cortana had begun as buggy and slow when competing against keyboard and mouse, and reviewers continued to characterize Cortana as markedly inferior to Apple’s Siri.
  • Forbes cited the Edge browser as “Microsoft’s attempt to claw back momentum from [Google’s] Chrome.” That helped to explain how Microsoft could benefit from pushing users toward Windows 10 — if, that is, Edge ceased to be “feature limited.” It did not explain how users would benefit — especially since, as with Cortana, there were concerns about Microsoft’s uploading of users’ private Edge search data.
  • Forbes said that the virtual desktops feature would give Windows 10 users a capability “long seen on Linux and Mac OS X.” Here, again, was a technology already found in Windows 7, and capable of being expanded and improved there.
  • Forbes attempted to portray the ability of Windows 10 to run on old hardware as a strength, but the more accurate phrasing was that Windows 10 required at least as much of a hardware investment as Windows 7, and that (as detailed below) it meant that Windows 7 users would be unable to use the newest CPUs from Intel and other manufacturers.
  • Forbes identified certain security advantages in Windows 10. InfoWorld specified that these included the removal of the insecure Internet Explorer as the default browser, the introduction of Windows Hello recognition of face, iris, or fingerprint authentication, and Device Guard. Microsoft‘s writeup of these features did not seem to explain why Windows 7 could not implement these features. The serious privacy breaches built into Windows 10 seemed likely to outweigh the value of these security advantages for most users on a day-to-day basis.
  • InfoWorld noted that Windows 10 added an Action Center — essentially, a tool that would manage system messages and notifications from various applications, with duplicative links to certain Control Panel features.Expressing the mixed response to the new Action Center, AskVG offered ways to disable it.
  • Windows 10 included were many other small fixes and advances. For instance, TrustedReviews (Chester, 2016) reported that Windows Explorer had been replaced with what the reviewer considered an improved File Explorer; window management was modestly improved; the Storage Spaces feature added the ability to group multiple hard drives together to form a single logical drive; there was enhanced support for multiple monitors; and gamers would appreciate the ability to access their Xbox Live accounts.

Computers: A Rumored Death Greatly Exaggerated

In a number of the industry analyses and commentaries I encountered during this research, there seemed to be a pervasive belief that the day of the (desktop or laptop) computer was past — that, if the traditional computer was not dead, it was dying. For instance, Barrons (Ray, 2015) described Windows as “a has-been operating system that is riding a steadily declining hardware platform, the PC.”

It was not clear what such writers could be thinking. Gartner expected that, in 2016, 290 million PCs (including desktops, notebooks, and premium ultramobiles) would be shipped worldwide. These would be added to the estimated total of 4.2 billion desktop and notebook computers shipped since the dawn of computing, of which an unknown but substantial number remained in use. Business Insider reported that, as of yearend 2011, Microsoft claimed that Windows was running on 1.25 billion PCs worldwide.

Gartner projected that, by 2018, PC shipments would rise to more than 300 million units annually. It was true that the rise would be slight, amounting to only about two percent per year. But with well over a quarter-billion units shipping per year, rising annual figures, and an enormous total installed base, the PC did not appear to be “declining.” And with Windows remaining dominant on PCs, and Gartner’s projection that adoption of Windows 10 would “boost the PC market in 2017,” it was rather ludicrous to describe Windows as a “has-been operating system.”

The concept appeared to be that smartphones and tablets had taken over the show. It was true that many more mobile devices than PCs were being sold. Gartner anticipated shipments of 2.4 billion mobile phones in 2016 — but those were expected to rise thereafter at only about 1.5% per year, and not all were smartphones. In fact, the mobile market appeared increasingly saturated. WeAreSocial (2016, pp. 499, 509) found that, in the U.S., the number of mobile subscriptions was rising only 4% per year. The situation was even worse elsewhere. In China, 99% of adults owned a mobile phone, and mobile subscriptions were rising at only 2% annually (pp. 128-129).

And then there was the question of what people were using these devices for. ExtremeTech (Anthony, 2012) maintained that, “In a few years, everything you do on your laptop today will be achievable on a smartphone.” But what actually happened was that, in a few years, USA Today (Saltzman, 2015) was still able to list only a few technical advantages enjoyed by smartphones and tablets vis-à-vis laptop or desktop computers — and even some of those (e.g., voice interface, GPS, reading ebooks) were available, sometimes in superior form, on computers. Architects, lawyers, video editors, and many others continued to find uses for bigger hardware based very much on a stationary surface. At this writing, for example, Amazon offered more than 600 different large-screen monitors ranging from 50 inches (that’s more than four feet, or about 1.5 meters) to more than 80 inches, at prices up to $100,000; and a number of the most highly rated gaming computers were bulky desktop machines ranging in price from $2,000 to $5,000 and beyond (e.g., Tom’s Guide, PCMag, EGCAmazon).

Without denying the utility of compact hardware in many settings, the ordinary situation in most homes and workplaces continued to be that people wanted to be able to type, use the mouse, see what they were doing on a screen large enough to be easily read, and enjoy powerful processing of their input. It was not surprising that, according to WeAreSocial (2016, p. 505), laptops and desktops accounted for two-thirds of all web pageviews. There also did not seem to be many people doing corporate accounting, writing books and research reports, running statistical analyses, or controlling nuclear power plants on smartphones. Generally, Ormond Beach observed that, compared to smartphones, desktop computers continued to be cheaper to maintain; were more powerful, scalable, versatile, and DIY-compatible; offered superior input options, larger displays, and more storage space; and were better for multitasking and virtualization.

The situation appeared to be, not that PCs were in decline in any meaningful sense, but that they were taken for granted. That was the mistake Microsoft had made with Windows 8: it had supposed that the touchscreen would take over, and therefore had stupidly disregarded the preferences of the familiar mouse and keyboard that enable so many people to achieve so much (see e.g., Scientific American, 2013; How-To Geek, 2015). Microsoft had still not sorted this out at the time of Windows 10 (DigitalTrends, 2016). So, for example, when Forbes cited a Microsoft representative for the claim that the Continuum feature in Windows 10 would allow a traveling user to use a smartphone or tablet as if it were a PC, BetaNews (Fagioli, 2015) pointed out that it might be more practical, for a real-life user in a hotel room, to have a regular keyboard and the other advantages of a traditional laptop.

For the foreseeable future, the world’s dominant computer operating system would continue to exert a profound influence on what humanity produced, performed, learned, and shared. Until there was a practical replacement for the PC, it would be bizarre to impugn it as a relic, or to treat PC operating systems (notably including Windows) as anything less than essential components in the development and growth of the contemporary world.

The Hard Sell

The slow pace of Windows 10 adoption was particularly remarkable in light of the pressure that Microsoft was placing on manufacturers, merchants, and consumers. This hard push included these measures:

  • HackerNews and Computerworld reported that, in the early weeks after release, Microsoft automatically downloaded Windows 10 installation files, of up to 6GB in size, without user consent, in some cases using up consumers’ allotted bandwidth. On computers where users did not retain control of update installation, Windows 10 installation was then reported to commence automatically. (Note the suggestions, below, on preventing such behavior.) Business Insider reported that downloads and installations without user permission were continuing into early 2016.
  • Unwanted Windows 10 upgrades apparently also occurred where users failed to notice that Windows 10 was being pre-checked as a supposedly optional upgrade. Some users reportedly found that the Windows 10 installation files were downloaded even after they had opted out, or that Windows 10 installation was offered as a fix to what appeared to be a system crash.
  • With behavior that many described as malware-like, the Windows Update service began to point more emphatically toward the Windows 10 upgrade, blocking other updates and facilitating accidental or unintended upgrades. It was possible to hide that Windows Update reminder, but doing so did not prevent repeated pop-up desktop reminders that Windows 10 was available as a free download and that millions of people had already upgraded.
  • Reputable commentators noted that Microsoft had suddenly started making Windows 7 updates dramatically slower, sometimes taking hours for a process that should have completed in minutes.
  • Computerworld (Keiser, 2016) reported that, regardless of the dates specified in its support lifecycle documentation, Microsoft had recently decided to reduce support for Windows 7 on the newest Intel CPUs. As MakeUseOf explained, due to pressure from Microsoft, Intel’s new Kaby Lake processors would not run anything older than Windows 10, and Microsoft support for Windows 7 on the previous Intel chip, known as Skylake, would end in 2017. The actual Microsoft announcement warned that Windows 10 would also be the only supported Windows on other chips — “for example, . . . Qualcomm’s upcoming ‘8996’ silicon, and AMD’s upcoming ‘Bristol Ridge’ silicon.”

AnandTech offered a mild characterization of likely corporate reactions to that last point:

For the average consumer buying a new PC, this is not a huge issue. . . . But the enterprise schedule is often much more drawn out when it comes to desktop operating system support. . . .

The move to Windows 7 was very drawn out, so perhaps Microsoft is trying to avoid this again in the future, but moving an enterprise to a new desktop OS can bring a lot of testing requirements, training, and back-end infrastructure updates which are all non-trivial. Microsoft has made its name in the enterprise by being generous with support lifetimes, and I think what is most troubling about today’s news is that Windows 7 has long-term support until January 14, 2020, and Windows 8.1 until January 10, 2023. News like this is going to catch a lot of companies off-guard, since they would have been expecting to have at least until 2020 to migrate off of Windows 7, and many of these companies have just finally moved to Windows 7 after a decade or more on XP.

To give just 18 months with these support policies is likely not what companies want to hear. This doesn’t mean that Windows 7 will be end of life in July 2017, but if you can’t run it on new hardware, this is going to put a dent in device sales too. If companies are not ready to move to Windows 10, they may have to stick with older hardware.

In many of the webpages cited in this section, and in the preceding section summarizing negative responses to Windows 10, it was clear that many consumers and companies were experiencing more than mild irritation at the quality of Windows 10, and at Microsoft’s heavyhanded tactics.

The Microsoft Mentality

At this writing, Microsoft was widely believed to have charted a new course. It did not appear, however, that the company’s internal culture had yet been radically altered. For purposes of trying to figure out what might come next, it seemed advisable to consider the company’s past and present behavior.

There was, first, the matter of customer loyalty. Business writers (e.g., Yahoo! Finance; BloombergBusiness) considered customer loyalty an important part of the successes that made Apple more than twice as big as Microsoft in terms of market capitalization (Wikipedia, 2015). For example, in a customer loyalty survey reported by The Motley Fool (Neiger, 2014), Samsung scored 35, Apple scored 28, and Microsoft scored -8. BetaNews (Williams, 2015) reported that, in that same survey, 41% of Apple users reported being satisfied, versus only 19% of Microsoft users. On the BrandKeys 2015 list of leading companies in terms of customer loyalty, Apple came in third (for its smartphones) and fourth (for its tablets); Microsoft was not in the top 20. Argus Insights found that, in terms of customer delight, Microsoft closed the gap somewhat during 2015, but still remained well behind Apple.

There were reasons for Microsoft’s poor showing in the area of customer loyalty. Some of those reasons had to do with the company’s long history of abusive and sometimes illegal behavior, at the expense of customers and the public. For example, those with long memories could recall “DOS isn’t done until Lotus won’t run.” That slogan from the 1980s was among many alleged misdeeds unearthed by the Justice Department during its antitrust prosecution of Microsoft circa 2002. The slogan meant that Microsoft deliberately designed its DOS operating system so that the popular Lotus 1-2-3 spreadsheet program (competing against Microsoft Excel) would function poorly.

People may debate the authenticity of the “DOS isn’t done” slogan, but concerns about Microsoft’s behavior were warranted in light of the European Commission’s 2004 determination that Microsoft should pay a fine of almost $800 million for its anticompetitive practices, and the 2001 decision by the U.S. Court of Appeals for the District of Columbia finding Microsoft guilty of anticompetitive behavior.

The U.S. government’s litigation against Microsoft spanned most of the 1990s. Chan suggested that the enormously expensive litigation, and its negative outcomes, made Microsoft much less aggressive, and prompted Bill Gates to hand over the CEO job to Steve Ballmer in 2000. Slate (Yglesias, 2013) argued that the litigation had good results — that, among other things, without government intervention, Google and Apple would never have been able to thrive within the system of market control that Microsoft had arranged for itself.

Chan noted that, to some, the departure of Gates damaged the corporate environment, making Microsoft much less creative and responsive. But Chan also said, “Gates had come across as arrogant, evasive and sullen in video depositions shown during the trial.” What was perhaps most important about the “DOS isn’t done” slogan was not that it may have conveyed actual Microsoft policy, but rather that it reflected the attitude that people could expect from what the Los Angeles Times (Helm, 1998) described as “a rather mean-spirited company driven as much by the desire to destroy competition as the desire to innovate.”

Chan quoted one observer as crediting Ballmer with saving Microsoft from being split up by the U.S. Department of Justice, and as saying that he also “managed to turn the company brand from one of fear and hate around the world to one of statesmanship and general benign partnership.” The Linux community of the 2000s did not seem to see Microsoft as benign, however. For instance, Schestowitz (2009) contended that Microsoft continued to make Linux adoption as difficult as possible.

In short, Microsoft was an aggressive and innovative company that grew to dominance and then became arrogant and complacent, to the great detriment of computer users and the public. In products ranging from Lotus 1-2-3 to IBM’s OS/2 to Linux, Microsoft consistently sought to suppress appealing alternative software that it did not control. The antitrust litigation, and Microsoft’s 2000s doldrums, were consequences of its own excessively aggressive behavior.

In the spirit of “what goes around comes around,” it appears that the company’s dirty-tricks approach to competitors was how entrenched players within Microsoft learned to treat their own colleagues. Dick Brass, former Microsoft vice president, told the New York Times in 2010 that

the big established groups [within Microsoft, such as Windows and Office] are allowed to prey upon emerging teams, belittle their efforts, compete unfairly against them for resources, and over time hector them out of existence.

It would not be surprising that aggressive and even predatory behavior would flourish within Microsoft during Ballmer’s tenure as CEO, from 2000 to 2014. Ballmer himself was of that nature. In 2013, former Microsoft senior VP Joachim Kempin blamed Ballmer for driving out numerous successful executives whom he saw as potential competitors (see Rigby and Gothard). Even while arguing that Ballmer was not one of the “five reasons why Microsoft can’t compete,” Wilcox (2010) agreed that the company had become a place of middle-aged lawyers and middle managers, protecting Office and Windows while leaving the riskier and more profitable opportunities to Google and others. Ultimately, the CEO was responsible for changing or else perpetuating that sort of culture. Kempin explained the dramatic fall in Microsoft’s stock price that many observers (e.g., EichenwaldThompsonZolman) attributed to Ballmer’s tenure as CEO:

Microsoft’s board is a lame duck board, has been forever. They hire people to help them administer the company, but not to lead the company. That’s the problem. They need somebody maybe 35-40 years old, a younger person who understands the Facebook generation and this mobile community. They don’t need this guy on stage with this fierce, aggressive look, announcing the next version of Windows and thinking he can score with that.

But it wasn’t just Ballmer. The form and marketing of Windows 10, arriving more than a year after Ballmer’s departure, were very much products of the Microsoft corporate culture — of a mentality, that is, with a long history of devaluing customer preferences, harming the customer’s best interests, and thus failing to earn customer loyalty.

At the risk of oversimplification, Microsoft had two options with Windows 10. It could do what it did: decide what customers were going to get, and ram it down their throats. Or it could start with where customers were: offer Windows 7.1 as a transitional step on the way to Windows 10, or design Windows 10 as a direct child of Windows 7, with obvious concern for enabling users to continue to use the hardware and software they needed. Microsoft was simply not committed to the goal of providing great computing.

In Windows 8, Microsoft tried to blend the computing needs of Windows 7 users with the touchscreen preferences of tablet and smartphone users. From the computer user’s perspective, this amounted to a loss of focus. It made Windows 8 inferior to Windows 7 for computing purposes. Windows 10 repeated this error, this time to facilitate an even broader muddling of the keyboard-and-mouse computer with an entire spectrum of present and future mobile devices. User preferences were disregarded; users were treated as de facto adversaries, engaged in a struggle to resist Microsoft’s strong efforts to overrule what users considered most appropriate for their own needs.

The Pivot to the Cloud

In 2014, Ballmer’s replacement began to challenge at least some aspects of Microsoft’s internal culture. The New York Times (Manjoo, 2015) described the new situation as follows:

Just about everyone in the industry believes the smartphone will remain the dominant computing device for the foreseeable future. But second place looks entirely up for grabs . . . .

Under Satya Nadella, who became Microsoft’s chief executive early last year, Microsoft is embracing a fragmented vision of the future, in which no single device, or even a single category of devices, reigns supreme. The plan is a bit crazy and rife with internal and external tensions. That doesn’t mean it can’t work. The future is unpredictable, so why not try a bunch of good stuff and see what sticks? . . .

You could say Mr. Nadella is fostering competition between all of Microsoft’s divisions. You could also say he’s setting up the company as a circular firing squad. . . .

The Surface Book . . . is a genuinely fantastic machine . . . . This suggests the ultimate goal of Microsoft’s new devices: They spark excitement for Microsoft more generally. . . . [T]he future is wide open. Microsoft is taking the first necessary step to tackle that future: Making really great devices.

That quote acknowledged that — as phrased by American Express (Hendricks, 2016), quoting management professor David King — “[T]he mobile model has disrupted people having to buy new software with a new PC.” The quote said, indeed, that, inside Microsoft, Windows was now competing for precedence against other initiatives, and to some extent had already fallen into a secondary role, with hardware (including gaming as well as mobile devices) being viewed as a key driver of the company’s future.

The quote also acknowledged that Microsoft was not a significant player in the realm of the smartphone, the dominant computing device. According to InformationWeek (Endler, 2014), Nadella initially sketched out a “mobile-first, cloud-first” strategy that would seek to make Windows the dominant operating system on mobile as well as non-mobile devices. A year later, unfortunately, Microsoft wrote off billions of dollars, and thousands of employees, in recognition that Ballmer’s acquisition of Nokia had failed. The New York Times (Stewart, 2015) agreed that Apple and Google had also experienced reverses in the smartphone arena, and noted that Microsoft did plan to move ahead with its Lumia phone in niche markets — but also remarked that Microsoft had also failed, so far, in competition with Apple in the area of wristwatch mobile computing devices.

Observers concluded that Nadella was seeking to insure that Windows would remain the dominant operating system by “increasing the reach of Windows 10 with every cloned device sold” (Weinberger) and “putting Microsoft solutions into the hands of as many mobile users as possible, in most any way possible” (Brugger). Microsoft’s strategy appeared to be to make Windows 10 the dominant operating system in as many kinds of devices as possible, grabbing market share — at the expense of short-term profitability, if necessary.

An important example: the company was giving its Office software, for free, to individual (as distinct from business) users of Apple and Android mobile devices. The Verge (Warren, 2014) noted that many would consider this crazy — after all, could it not encourage consumers to buy an iPad rather than a Microsoft Surface? — but that at least it would discourage competitors from advancing Office alternatives in the mobile space. In another move that Microsoft stalwarts (including Ballmer) always rejected, Wired (Metz, 2014 1 & 2) described why Nadella opened Microsoft’s Azure cloud computing service to support Linux distributions and virtual machines, and also open-sourced many of its software tools and programming languages:

[Nadella] realizes that in today’s world, Microsoft must operate more like Google. You don’t succeed by trying to force an expensive operating system onto the market. You expand your tech empire by offering free operating systems and free developer tools. You can then make your money by selling other stuff, like web services and online ads and maybe even Microsoft Office. . . .

[B]y open sourcing its developer tools . . . Microsoft can get them into the hands of more coders, and that means these coders will build more stuff for Windows and other Microsoft platforms — at least in theory. . . .

[Ex-Microsoft executive Sam] Ramji agrees that free .NET is the way forward . . . [but] he thinks the change may have come too late. Because programming tools like Java have been open source for so long, they’re already blanketing the programming world, and things like .NET have an awful lot of catching up to do.

In the words of ZDNet (Vaughan-Nichols, 2014), “Microsoft’s fortunes now lie not with the desktop or desktop programs, but with its Azure cloud and cloud-based programs such as Office 365.” Azure had to offer Linux services, keeping up with its competitors, if it hoped to survive.

Why Did Microsoft Need Windows 10?

From a computer user’s perspective, Windows 10 was a solution in search of a problem. This operating system upgrade was driven, not by user demand for new functionality, but by Microsoft corporate imperatives. As shown above, those imperatives had to do with mobile and cloud computing. But why, exactly, did Microsoft find it necessary to push a new Windows operating system?

A search led to a video suggesting that Microsoft was motivated primarily by a desire to emulate Apple’s success in linking its various devices, from smartphones to laptops, via a single operating system. This emulation would include not only the OS itself, but also a single online store offering apps that would run on all Microsoft devices. In this theory, the upgrade to Windows 10 was free because Apple likewise offered free upgrades to its operating system. The video also indicated that Microsoft was reacting against Google, whose various offerings had moved traditional operating system functions to the cloud. Microsoft’s concern in that direction was to demonstrate that an operating system (Windows, in particular) can deliver much more than cloud-based computing can. But Windows 10 was not coming across as a winning alternative to the cloud, and there were limits to how much Microsoft would want to position itself as an opponent of cloud-based computing, given its earnest efforts to gain market share for its Azure cloud system (below).

V3 (Worth, 2015) offered a somewhat different hypothesis. The claim here was that Microsoft realized that it was very far behind in the mobile device battle, but was looking ahead to the Internet of Things (IoT). The IoT consisted of networked devices that would collect, transmit, exchange, and use data, from machine to machine (M2M), and between sensors and machines, largely without human involvement (Wikipedia; Wired 1 & 2; Gartner 1, 2, 3). Examples of such devices ranged from smart homes to smart cars to smart toothbrushes to sensors in the floors and highways to biochip transponders on cows — altogether, at least 20 (by one recent estimate, 40) billion devices autonomously feeding information to each other and/or to humans by 2020, and raising huge privacy and security concerns along the way. The hypothesis was that these devices, running Windows 10 IoT, would use Azure to tie together the machines and the data in coherent and useful ways. V3 quoted one analyst as saying, “Microsoft has not often been seen as an innovator in the past few years but with Nadella at the helm they are definitely a leader in this area.”

In IoT computing, Microsoft would face competition, not only from Google and Apple, but also from AmazonIntel, Qualcomm, Samsung, and others. Writing nearly a month after the launch of Windows 10, TechRadar (Carter, 2015) favored Google and Qualcomm, especially within the home as distinct from industrial settings — and did not even mention Microsoft as a contender. Such subpar visibility may have prompted Microsoft to join the AllSeen Alliance and, perhaps more importantly, the Open Connectivity Foundation, whose members included Intel, Cisco, Samsung, IBM, Dell — indeed, almost everyone, it would seem, except notably not Apple, Google, or Amazon. The Foundation’s stated purpose was to facilitate communication among IoT devices “regardless of manufacturer, operating system, chipset or physical transport.” Its unstated purposes no doubt included a desire to give these corporations a fighting chance in a mobile market (and perhaps already, to some extent, an IoT) dominated by Apple and Google.

The stakes were very large. There was the possibility that, someday, a company achieving IoT supremacy would be closer to controlling the world than any organization or government in history. It appeared, then, that the need for action was urgent — that Microsoft rushed Windows 10 out the door as soon as it could offer at least a moderately credible foundation for work on devices ranging from established computers and smartphones to the emerging and still somewhat unknown devices of the future IoT.

The argument advanced here, then, was that Microsoft did not need Windows 10 for the desktop or laptop. It needed desktops and laptops to use Windows 10 so that Microsoft could offer, a straight face, “one core Windows that runs everywhere,” as Nadella described it (ZDNet, 2014). At this writing, recent comments (1, 2) suggested that the company had invested a great deal in putting Azure among the best cloud services, and thus appeared to be positioned to make inroads into the IoT market.

According to The Motley Fool (Kline, 2015), Microsoft’s strategy “to own the Internet of Things” depended upon attracting the attention of developers. Developers were needed to create the software applications that would run IoT devices and link them with other devices. To that end, Fool said, “Microsoft has done some of the heavy lifting for developers” by replacing the former “closed-off world” of proprietary Microsoft code with a new approach “generally inviting everyone to play with its OS.” As explained by Forbes (Kelly, 2015), developer acceptance of Windows 10 would result in more Microsoft-compatible universal apps; and the existence of more apps, meeting more user needs, would make the company’s smartphone, tablet, and gaming hardware more appealing to consumers. Developers, wishing to write programs that sell, would understandably focus their efforts on operating systems that would help them to reach large markets. Microsoft seemed to be gaining momentum through purchases of some of the best apps on Google and Apple hardware — through, more generally, the impression that “Microsoft has become a progressive, open, web services company” while “Google is letting the basics slip” (Kelly, 2015).

Microsoft predicted that Windows 10 would be installed on a billion devices by 2018. Notably, that prediction included IoT devices. That inclusion made the target easy, as long as Microsoft achieved at least modest IoT market penetration. To accomplish that, Microsoft did not need a vast and prompt shift away from Windows 7 on PCs. Massive success for Windows 10 would have been nice, but it was not worth additional months of delay, just to make sure of the quality of Windows 10. What Microsoft needed from Win10 was just enough of a step forward so that people who were making important decisions (most of whom were Windows users) could comfortably choose Windows-based products for their own upcoming IoT purchases.

(After completing this post, I came across an article suggesting that the push toward Windows 10 was also motivated by a desire to diminish Windows piracy in China. This did not immediately seem to be a compelling motivation. Microsoft had been treating desktop computing as something of a relic ever since the introduction of the Windows 8 interface. The article indicated that, whatever Microsoft might offer, Chinese users were not feeling particularly eager to switch away from Windows 7. And if they were, apparently they are increasingly tempted to switch to a home-grown Windows alternative. In China, as in the U.S., it seemed that Microsoft may have shot itself in the foot.)

Microsoft’s New Priorities

Summarizing and extending some points discussed above, Directions on Microsoft (Helm, 2016) offered a succinct statement of Microsoft’s priorities for 2016. Key excerpts:

Microsoft-hosted cloud services and products for mobile devices will continue to dominate the company’s enterprise strategy . . . [with a] focus on platform: Windows Server, SQL Server, Azure services, and Windows 10. . . .

[In the cloud, among other things,] Azure services enable customers to rent virtual machine computing, storage, and networking infrastructure and to run server applications . . . . Office 365 services offer e-mail and other collaboration and business intelligence (BI) applications. . . .

[Mobile offerings include] software such as Office for phones, tablets, and other mobile devices . . . . Microsoft mobile software and services have recently focused on the iOS and Android OSs that dominate the market. However, Microsoft continues to improve Windows as a mobile OS and promote it through Microsoft-branded mobile devices such as Surface tablets and Lumia phones . . . .

Both cloud services and mobile software are sold on a subscription basis, as Microsoft tries to shift from the less predictable business of selling perpetual licenses. Cloud and mobile businesses have also encouraged the company to release and retire software more frequently . . . .

By expanding its cloud and mobile businesses, Microsoft has built alternatives to its main PC software businesses, the Windows client OS and the Office desktop suite, which have suffered from slowing PC hardware sales. . . .

The Microsoft-developed Surface line continues to expand . . . . The hardware business enables Microsoft to promote Windows 10, exploit new component technologies (and competitor mistakes), and develop new types of devices such as the HoloLens augmented reality headset . . . . Consequently, the company will stay in the hardware business for the foreseeable future. . . .

With more frequent upgrades come more frequent Microsoft product support deadlines . . . . For systems that need to be kept stable, however, customers need to select specific software editions (such as Windows 10 Enterprise) that are updated less frequently, use administrative tools to opt out of Feature Upgrades until they have had time to test them, and ensure that computers get software from organization-controlled sources rather than from Microsoft directly. The faster pace of releases also means that customers might need to test each release more thoroughly than in the past.

PC users might be dismayed by several points conveyed in those excerpts and the preceding discussion. First, Microsoft prioritized Windows 10 as a tool for access to the IoT — not as a superior solution for desktop and laptop computing. Indeed, the company saw its cloud and mobile business as “alternatives” to Windows and the Office suite. In the cloud and mobile areas, and perhaps also on the desktop, Microsoft was moving toward software subscriptions that would generate steady income. Programs would be added and retired more frequently. Microsoft would be selling its own computers and perhaps moving toward the Apple model, where at some point those who wanted to use Windows might have to buy Microsoft hardware to do so. (By contrast — to anticipate the alternative discussed in Part Two (below) — The Inquirer (Merriman, 2016) reported that Linux was compatible with the newest Intel CPUs.)

Where’s the Money?

Many wondered how Microsoft would make money, if it gave away Windows 10. But there seemed to be multiple possibilities. For one thing, the free Windows 10 upgrade was only available to people who already had Windows 7 or 8. First-time buyers and people migrating from other versions of Windows (e.g., XP) would still pay. Also, OEMs would still be charged for copies of Windows they would install on computers sold to end users. Microsoft could also charge for future Windows upgrades (to e.g., Windows 10.2). So it was not as though Microsoft was foregoing all revenues from sales of Windows 10. PCWorld (Chacos, 2015; see also Phillips, 2015) pointed out that Microsoft also made money from the use of Windows 10:

Cortana ramps up Bing’s market share with every search you make. OneDrive backs up everything to the cloud, and of course you can buy more storage space if you need it. The Video, Groove Music, and Xbox apps encourage entertainment purchases through Microsoft. The new Edge browser and the very operating system itself track you to serve targeted ads. The free Office apps encourage paid Office 365 subscriptions to unlock full functionality. Underneath it all, the Windows Store is the repository for all of Microsoft’s vaunted universal apps (and plenty of other things to buy). Heck, even Solitaire begs for a monthly subscription to ditch ads now.

But perhaps the more important point was summarized by ZDNet (Bott, 2015): “Big companies are the core of Microsoft’s business. . . . [R]etail sales of Windows, both full copies and upgrades [are] a microscopic business.” As noted above, forced updates were a point of complaint by those who did upgrade to Windows 10. Windows 7 and 8 users would pay for their free upgrades to Windows 10 by having updates forced upon their systems, with the risk that some of those updates would cost them time, money, and/or data.

In effect, Microsoft was using these free upgraders as beta testers — “guinea pigs,” in the term suggested by Forbes (Kelly, 2015) — for the big companies that provided the bulk of Microsoft’s revenues. In Microsoft‘s phrasing, “[E]nterprises will be able to receive feature updates after their quality and application compatibility has been assessed in the consumer market.” Experimentation at the expense of millions of end users would provide assurance that corporate upgrades and updates would be stable and well-tested in the field. This may explain why free upgrades were not offered to users of Windows XP and other older versions: corporations would rarely be upgrading from those old versions, so those users’ upgrade experiences would not be especially helpful for purposes of making Windows 10 ready for corporate customers. (The option of getting a safer version of Windows 10, such as the Enterprise version recommended at the end of the previous section, is discussed below.)

What Could Happen Next

The foregoing sections of this Part One suggest several possible future developments. This section reviews that information and sets the stage for Part Two, below.

What could and probably would happen next, as of this writing, was that Microsoft would proceed with the various initiatives described above, regarding the cloud and mobile computing and other corporate objectives. The concern here was, more specifically, what might happen to the use of Windows 7 on PCs.

One thing that could happen next was that Microsoft could intensify its efforts to undermine Windows 7, so as to drive the majority of Windows users on to Windows 10 as soon as possible. In the transition to Windows 10, as noted above, Microsoft had already shortened its support period for Windows 7 and insured that Windows 7 would not run on the latest CPUs. Some felt, moreover, that Microsoft was engaging in “deliberate misinformation designed to unsettle users” (Kelly, 2016).

It also appeared possible that, by commission or omission, Microsoft might sabotage Windows 7. Doing so by commission would mean taking deliberate actions intended to make Windows 7 systems unstable and/or to make a Windows 10 installation inevitable. Along those lines, an Overclock thread (2016) contained a list of Windows 7 and 8 updates that seemed designed to prepare systems for Windows 10 installation. Sabotaging Windows 7 by omission would mean failing to take actions needed to insure system stability. At least some omission was guaranteed: at a certain date, Microsoft was no longer going to maintain Windows 7. More ominously, however, ZDNet (Gewirtz, 2015) noted that his Windows 7 installations had been “rock-solid until Microsoft started to harass us about upgrading . . . to Windows 10.” Gewirtz considered that Microsoft might well go “out of its way to break [Windows 7], in a misguided attempt to convince us all to move to its new version.”

Deliberate sabotage might lead to lawsuits. But deliberate omission — failing to allocate sufficient corporate resources to insure that future updates ran properly, for example — could be a different matter. Without technical familiarity with the inner workings of Windows code, one could not be entirely confident about what might already be in there. A piece of code that would have to be periodically updated to continue to run properly, for instance, could become the first victim of update neglect, even before the official end of Microsoft service.

Did Windows contain such built-in failure zones, or would Microsoft distribute updates, in 2015 or 2016, that would increase the likelihood of Windows 7 failure? It was not reassuring that the company did have that alleged history (above) of undermining programs, like Lotus 1-2-3, that it wished to put out of commission.

I had been tempted, in the past, to suspect sabotage when an existing version of Windows became unstable, just when Microsoft was promoting some new version. I had found it difficult to understand why an image backup of my drive C, made when the system had been performing stably, would become unstable when restored to the system at a later date. Likewise, others (e.g., 1 2 3 4 5 6) reported their own experiences, giving rise to concerns that Microsoft might indeed be taking overly aggressive and even deliberately malicious steps to promote Windows 10 at the expense of Windows 7.

Perhaps the central concern here was that Microsoft appeared to be ending the basic formula that had been in place for decades: one buys a computer, installs some software, and goes to work. Now it was more complicated, and it was all subject to change within a relatively short time: the computer might no longer be sufficient for the software, the software might not function like it used to, and so forth (see Paul, 2015).

In response to those concerns, another possibility (though seemingly unlikely) was that Microsoft was tightening the screws on Windows 7 users to make them motivated buyers of something like Windows 10 Enterprise edition. Microsoft’s message here could be, we will offer you a way to continue using something very like Windows 7, without the disruption likely to be experienced by the general computing public using Windows 10 as de facto beta testers — but it will cost you.

The End of the Free Upgrade

More immediately, there was the question of what would happen to the Windows 10 free upgrade offer, on its one-year anniversary date in July 2016. Commentators appeared to be leaning toward the belief that the offer would be extended. For example, ZDNet (Bott, 2016) speculated that Microsoft would extend the offer for one more year. That speculation was based on the assumption that Microsoft would not begin to charge for Windows 10 upgrades in July 2016 because doing so would discourage Windows 7 users from upgrading. Microsoft supposedly wanted them to upgrade in order to avoid “an XP-style mess.”

The “XP-style mess” seemed to be a reference to what Forbes (Kelly, 2015) described as “13 years of free development and tech support followed by customer scorn when [Microsoft] eventually called time.” That historical experience did not seem applicable in the present setting. It was clear that Microsoft’s support for Windows 7 was going to terminate at a specific date. What seemed most likely to invite score was, ironically, that Microsoft seemed to have grown hardened to scorn: it seemed intent on ramming Windows 10 down the throats of unwilling users much more energetically than before. It is doubtful that Nadella had any illusions that there would remain a residual base of users of Windows 8, Windows 7, and other previous versions, or that many users would continue to be displeased with Microsoft’s behavior.

ZDNet argued that Microsoft would not extend the free upgrade offer indefinitely, because doing so would “provide an excuse for laggards to keep kicking the can down the road for as long as possible.” But the same could be said for a one-year extension to July 2017. To the extent that ZDNet was correct in identifying a Microsoft desire to wrap up Windows 7 as thoroughly and quickly as possible, it would seem that Microsoft would want to increase, not reduce, the pressure favoring a Windows 10 upgrade.

There was some plausibility in ZDNet’s contention that “Asking existing Windows 7 users to pay $99 or more after they’ve spent a year avoiding the free upgrade seems like a surefire way to guarantee that they never upgrade.” But that would not be the only option. Microsoft could announce, in May 2016, that it would be making Windows 10 available to Windows 7 users, on a limited-time basis, at a price of $29. This approach would clarify that the free offer was definitely going to end. It would also establish that the price was going to be rising. And the price could indeed rise again, after six months or a year.


Part One (above) contains information pointing toward several possible responses to what was, at this writing, an offer for a free upgrade from Windows 7 or 8 to Windows 10. This Part Two discusses those possibilities.

Short-Term Fixes

It seemed that the most immediate practical response, to the free Windows 10 upgrade offer, would be to acquire imaging software and make a backup image of one’s Windows installation on drive C. This was practical, regardless of Windows 10, as a way of protecting oneself against loss, damage, or instability in one’s Windows installation. As I had found on many occasions, an image would convert a multiday Windows reinstallation project into an image restore process taking less than an hour.

With a Windows 7 image safely stored on a backup drive, another practical step would be to temporarily install Windows 10. For people who were not currently running Windows 7 or 8, this might require first obtaining a cheap copy of Windows 7. ZDNet (Bott, 2016) offered a list of ways to do that: buy a new or used PC with Windows 7 preinstalled; buy an OEM copy of Windows 7; downgrade for free from Windows 8; transfer a Windows 7 license from another computer; buy a Windows 7 upgrade license; or get a copy from your educational institution. MakeUseOf (Phillips, 2015) added ways to buy just a Windows 7 license key (i.e., Bonanza, G2A, Amazon, MicrosoftSoftwareSwap, PriceGrabber, but not fake shops).

My other post describes what I did to install Windows 10 on a Windows 7 machine. The objective would be to get Windows 10 running and activated. This would give the user the desired free update. There was no rule requiring him/her to use it. Once that was done, the user could use the imaging software to make a backup of the ready-to-go Windows 10 installation. With the Win10 image safely stored and ready for future restoration to the hard drive as needed, the user could restore his/her most recent Win7 image and go back to work.

Before and/or after making an image of the Windows 7 drive, the user might want to follow advice on how to prevent Microsoft’s pushy Windows 10 reminders and unwanted downloads. That advice included the following:

  • Leonhard suggested using GWX Control Panel — but he noted that this might not work for users who had already gone too far in the Windows 10 download process. At that point, he said, only a System Restore to a previous date might work, though it might also introduce additional hassles. (Presumably restoring a previous drive image would also solve the problem.) A more recent and perhaps better solution: Never 10 from Gibson Research.
  • Alternately, there were manual approaches. For instance, Minasians explained that the upgrade nagware was installed via a Windows 7 and 8 upgrade known as KB3035583. If the update was already installed, the first step was to go into Control Panel > Programs and Features > View Installed Updates > click on Name column heading to sort alphabetically > wait for the sort to finish > scroll down to Update for Microsoft Windows (KB3035583) > right-click > Uninstall > restart the computer > go back into Windows Update. At that point, regardless of whether the update had been previously installed, the advice was essentially to see whether Windows Update indicated that an “important update” was available. If not, the user could click Check for Updates. Whenever KB3035583 did once again appear as an available update, the next step was to right-click on the KB3035583 update > Hide update > close the updates window (or, if applicable, install other updates as desired). My first effort to apply this fix resulted in a mouse pointer on a blank black screen after rebooting, which I was only able to resolve by restoring my most recent Acronis drive image. For some reason, it worked OK on my second try.
  • Microsoft (see also ZDNet, Hacker News, PCWorldComputerworld) offered several methods to do several related things in Windows 7 and 8. In Win7, one method was to copy this information into Notepad, save it as a .reg file with a name like PreventWin10Upgrade.reg, and run it:
Windows Registry Editor Version 5.00

; Block the upgrade to Windows 10 through Windows Update

; Hide the Get Windows 10 notification icon
; Previous method of hiding the Get Windows 10 (GWX) icon
  • WindowsCentral advised looking for the hidden folder C:\$Windows.~BT. (To see hidden folders, go to Windows Explorer > Tools menu pick > Folder Options > View tab > Show hidden files, folders, and drives.) If the folder was there, reportedly this would indicate that the 3-6GB Windows 10 installation files had already been downloaded onto the local drive. WindowsClub said there could also be a $Windows.~WS folder, at least for those who had already upgraded to Windows 10, and that these could be deleted by various commands if a simple Delete effort failed. For those who had not upgraded, it appeared that installing Unlocker and using its right-click context menu option in Windows Explorer might also be convenient for deleting that folder. AddictiveTips offered further informatio.
  • There was also the possibility of disabling the Windows Update service altogether. According to, this was a simple matter of going into services.msc and disabling the Windows Update entry. VMware offered a different approach for Windows running in virtual machines.

A Longer-Term Alternative:
Windows 10 Enterprise LTSB

Much of this post discusses the fact that, for reasons related to its internal culture and corporate objectives, Microsoft had repeatedly engaged in flaky behaviors disruptive of the user’s computing experience. The Windows 10 upgrade confusion constituted only the latest installment — the latest upgrade, one might say — in that long-running saga.

In the foregoing discussion, it appeared that Windows 10 Enterprise edition might emerge as a potential solution, for the Windows 7 user who might appreciate Windows 10 more if it would behave coherently. Microsoft explained that Windows 10 came in six editions: Home, Pro, Enterprise, Education, Mobile, and Mobile Enterprise (not mentioning IoT). Leaving aside the Education and Mobile editions, TechNet (1 & 2) described the update possibilities for these editions:

  • Windows 10 Home, Pro, and Enterprise could all participate in the Current Branch (CB) approach to updates. In the CB, updates would be downloaded as soon as Microsoft published them. This would have the advantage of making new features available immediately, but the drawback of making the CB user a guinea pig (above).
  • Windows 10 Pro and Enterprise (but not Home) could instead choose the Current Branch for Business (CBB), in which updates would be downloaded about four months after Microsoft published them, allowing time for the mass-market guinea pigs to try them out.
  • Windows 10 Enterprise would have the option of participating instead in the Long Term Servicing Branch (LTSB). LTSB offered “long-term deployment . . . in low-change configurations,” typically for critical or specialized devices. WindowsITPro (Trent, 2016) offered the examples of the ATM or the automobile assembly line: a bank or manufacturing company would not want these to come to a halt at random times in order to install an update and then reboot. LTSB did have characteristics that some would find appealing. For instance, most Windows 10 built-in apps (e.g., Cortana, Edge, Store) were removed from LTSB. According to WindowsCentral, the LTSB concept was that “some businesses don’t want to deal with testing new features that Microsoft might want to add to Windows 10.” LTSB would reportedly be supported for ten years.

It seemed, in other words, that Microsoft had reviewed its sources of income and had concluded that providing a stable computing experience was somewhat optional for its individual users but crucial for its enterprise customers. The Windows 7 user who looked for stability in Windows 10 would therefore tend to expect something like the CBB or LTSB approach to updates. And, indeed, there was some interest among Windows 7 users. For instance, Winaero concluded that “the LTSB edition is more like Windows 7” in terms of updates, and participants in an NTLite thread expressed interest in LTSB as a Windows 7 replacement.

Regardless of interest, it was not clear that individuals could acquire and afford LTSB. Microsoft’s Windows 10 for Enterprise page offered a link to a Buy Online option, but unfortunately that led to the Windows Store, which offered Windows 10 Home and Pro but not Enterprise. Another Microsoft page said that the user could “Get Windows 10 Enterprise edition when you purchase the Enterprise Cloud Suite (ECS) with Office 365 and the Enterprise Mobility Suite (EMS).” That sounded expensive. Following the links to those several products revealed that the cost for Office 365 would range between $5.00 and $12.50 per month, depending on version, and EMS would cost $8.75 per month. The link to ECS did not seem to work: it led to another page on EMS.

Eventually, I found a Microsoft page indicating that the Microsoft Enterprise Agreement was only available to organizations with at least 250 users or devices, and it looked like that might rise to a minimum of 500 users in 2016. The Verge and Winaero agreed that, ultimately, Enterprise was only available to institutions. Winaero did suggest that there might be illegal ways “to get Windows 10 LTSB and activate it.” Short of that, it seemed that the best a Windows 10 user could do, in search of a more controlled update experience, might be to opt for CBB. URTech offered an explanation of how to turn that on in an ordinary (e.g., Pro) copy of Windows 10.

Microsoft did offer a free 90-day evaluation copy of Enterprise. AskVG offered a product key for that evaluation copy, though apparently no such key would actually be required. AskVG suggested installing the trial version “in a virtualization software such as Virtual Box or built-in Hyper-V feature.”

At this writing, then, it appeared that Windows 10 Enterprise LTSB could be of interest to individual Windows 7 users as “the most lightweight edition of Windows 10” (AskVG), but that Microsoft was not presently set up to sell that edition to individual end users.

Non-Windows Alternatives

It did not presently appear that competitors were making major inroads into Microsoft’s corporate operating system sales. ComputerworldUK (Finnegan, 2016) observed that “Apple is continuing to gain a foothold in the enterprise at the top-end, while Chromebooks have offered a cost-effective alternative to Windows,” but that “Most organizations still have a dependency on Windows.”

Much the same seemed to be true at the individual level. Forbes (Kelly, 2015) summarized the situation as follows:

Windows users are essentially a captive audience. After all Windows 7 and Windows 8 users can voice their displeasure as much as they like at these upgrade tactics (and many are), but what are they going to do about it? The experienced may switch to a Linux variant and learn how it works while those with the cash to afford it might migrate to Apple and buy all new hardware and software, but for most these simply aren’t credible options.

As such Microsoft can essentially do what it likes. The vast majority of its user base will stay put regardless of what the company does and however devious it becomes in forcing/tricking users to upgrade. Moreover Windows 10 is actually a good (if Orwellian) operating system and I’m sure the company believes that those fighting the switch will calm down once they experience the upgrades Microsoft has made.

In a comparison of Mac and Windows, ZDNet (Gewirtz, 2016) concluded that each had its strengths. Cost was indeed a factor in favor of Windows, and so were the variety of hardware options and gaming, while Mac came out ahead on consistency of quality and on security. In terms of variety and flexibility, it seemed that Windows power users might find the Mac environment too limiting. There was also the observation that Mac OS X would feel weird to a Windows user. Informatics-Tech suggested that (1) Windows enjoyed the advantage in compatibility (“Almost every application, driver or game will work on Windows”), gaming, tech support (there was almost always someone, somewhere, with an idea of how to address a problem), and functionality (there was almost always a way to do something), as well as the disadvantages of being prone to viruses and slowed by heavy demands on computer resources; and (2) Mac OS X was less prone to crashes and had a better appearance, but was more expensive than Windows, and required Mac hardware and software. Other comparisons seemed to suggest that Windows was perhaps more competitive with Apple, on a variety of issues, than the foregoing Forbes quote indicated. Many Windows 7 users would be fairly heavily invested in the time and knowledge required to do things in familiar ways, such that the learning curve and lost time involved in a switch to Mac would be a significant barrier.

The learning curve would be less of a concern in a switch to Google’s Chromebooks, running the Linux-based Chrome operating system, because Chrome sought to simplify the user experience. (Wallen, 2015) reviewed several Chrome OS imitators that would enable users to run a Chrome-like environment on common x86/64 hardware. ArsTechnica (Cunningham, 2015) discussed the use of CloudReady to easily convert an old Windows PC into a Chromebook — with, unfortunately, some significant limits in what could be done (e.g., inability to play MP3 and MP4 files; no support for Netflix). That ArsTechnica article admitted that “CloudReady doesn’t quite elevate Chromium OS to the level of a Linux distribution that you could expect to install and run on any old system.” No doubt some Windows 7 users would find that a Chromebook (or an imitation thereof) would meet their needs. But then, so would Windows 10, most likely. For instance, Laptop (Palladino, 2015) concluded that Windows 10 laptops beat Chromebooks for most purposes. It appeared unlikely that a Chromebook would provide a tempting solution for most Windows 7 users whose particular circumstances and needs led them to reject Windows 10.

As another option, instead of going for Mac or Chromebooks, Linux offered a smorgasbord of possibilities (see Harvey, 2014), with a learning curve that might not be too steep for basic GUI functionality but that might increase notably with more esoteric uses. The well-known Ubuntu distribution was popular in itself and as the basis for other Linux tools, such as Cub Linux, which How-To Geek characterized as a Chrome-like desktop that incorporated standard Linux apps. Ubuntu had been superseded in popularity by the Ubuntu-based Linux Mint, whose Cinnamon edition was named as the best desktop Linux distribution for 2016. CIO (Bhartiya, 2014) confirmed that Mint’s Cinnamon would indeed be most familiar to Windows users. A review by (Bhartiya, 2015) raised the question of whether Mint or Ubuntu could be configured to offer a Start Menu similar to that of classic Windows. A search led to several demonstrations of Ubuntu and Mint installations customized to have more of a classic Windows appearance than Windows 10 could offer (see also NoobsLAB and UBTutorials).

There had been efforts to construct pre-packaged Windows-like Linux releases. TechRepublic (Sanders, 2016) characterized one such effort, ReactOS, as “open-source Windows” but noted that it was “still generally considered alpha-level software” (i.e., very much unready for prime time). For old machines, LXLE was an option. (As an aside, some users might find it useful to experiment with Ubuntu as a program running within Windows via Wubi.)

A different effort, the Ubuntu-based Zorin OS, was available in four versions, ranging in price from free to about $11. PCWorld (Noyes, 2014) offered “10 Reasons to Try Zorin OS 9.” (Wallen, 2016) reported that Zorin OS offered a computing experience “incredibly close to Windows XP.” The reviewer (2016) agreed that Zorin “behaves a lot like Windows.” TechRepublic (Wallen, 2014) had previously found Zorin very Windows-like, but lacking in the help department. Tech-FAQ (2015) discussed Zorin in an article titled “Promising, but Still Typically Linux.” The article explained that, in the author’s test run, there were a number of snags in installation and operation.

Generally, the Zorin effort seemed to be oriented primarily toward relatively superficial aspects of the operating system interface. For example, it did not appear that Zorin offered a Windows-like Command box; there also appeared to be no Control Panel or Device Manager. It seemed highly unlikely that such features would appear in a Linux distribution. The difficulty of developing such features in Linux may have explained why ReactOS was still alpha software. If it was a matter of getting the mere appearance of Windows, it made more sense to use a highly regarded and broadly supported mainstream Linux distro like Mint or Ubuntu, given the ability to configure them, too, to have a Windows-like appearance.

For users wanting more than a Chromebook could offer, then, Linux and Mac’s OS X operating system seemed to offer the primary non-Windows alternatives. At this point, having compared OS X against Windows, it made sense to compare OS X against Linux. A search identified a number of recent comparisons:

  • Datamation (Hartley, 2015) contended that Linux was more practical than OS X because Linux would run on a wide variety of hardware, made it easier to install the operating system onto a new drive, and was as easy to use for ordinary users. Datamation suggested that OS X would enjoy more local support, however.
  • InfoWorld (Lynch, 2015) offered replies to that Datamation piece by (1) a Mac user who contended that the smaller number of hardware options made troubleshooting easier, whereas some Linux issues were never resolved; (2) a Linux user who suggested that troubleshooting in Linux was actually easier, once you got the hang of it, because more information was available (due perhaps to the open source nature of the system); (3) commenters who alternately felt that Windows Explorer or Mac Finder could be incomprehensible to non-geeks; (4) someone who felt that limited support was precisely why Linux had not gone anywhere; and (5) commenters who disputed the Mac reputation for hardware quality, contended that Macs failed to perform as expected about as often as Windows machines did, and criticized the helpfulness of the Mac support community and documentation.
  • DZone (Seibert, 2013) saw it as a matter of the right tool for the job: a Windows machine for gaming, OS X at home, and Linux for software development, thanks to its superior automated installation of software and system upgrades.
  • ChannelPro (Jones, 2015) largely echoed, or slightly modified, the foregoing views, but stated that “If anything the transition from Windows to Linux is easier than Windows to Apple OS.”
  • ZDNet (Vaughan-Nichols, 2016) offered several opinions: unlike Windows, the Mint desktop kept getting better; Windows users “won’t have any trouble” with the user interface; Cinnamon 2.8 offered “the ultimate Window, Icon, Menu, Pointer (WIMP) interface . . . perfect for long-time PC users”; “you can just sit down and start opening directories, running applications, and modify your PC’s settings”; Mint offered multiple virtual desktop formats; Linux was far more stable than Windows or Mac; and Linux would always be free. On the other hand, the article said, “If you want one operating system family on all your devices, don’t waste your time — for now — on either Linux or Windows. Just go ahead and buy an iPhone and a Mac and be done with it.”

Those sources seemed to indicate that Mac might be best for certain needs. On balance, though, for a person coming from Windows (and with the benefit of prior experience with Linux and Mac), it appeared that Linux might be a better fit for someone who might consider him/herself a Windows 7 power user, or who might be looking for Windows-like flexibility, appearance, and functionality, without unnecessary expense and with minimal risk of corporate manipulation and disruption — bearing in mind the many possible criticisms of Linux at its worst.

Windows Programs or Linux Counterparts

The thought of replacing Windows with Linux immediately raised the question of whether one could find and use programs, in Linux, to accomplish familiar tasks in Windows. There were two ways to answer that question. This section takes a program-by-program approach, seeking ways of finding replacement Linux programs or of bringing Windows programs over to Linux. The next section considers the alternative of running the entire Windows operating system inside of, or in connection with, the Linux system.

There were no Linux versions of some of the best-known and most expensive Windows programs, such as those offered by Adobe (e.g., Acrobat, Photoshop). But there were Linux versions of many other Windows programs, including Firefox, Google Chrome, Google Earth, Skype, VLC Media Player, 7-Zip, and Picasa. The person who used such programs in Windows would typically face little to no learning curve when using the Linux versions of those programs. There was also the option of using cloud-based tools — which, again, would tend to work the same on any operating system. Examples included Google Docs, Google Maps, Netflix, Dropbox, and online tax-prep software.

Where there was no Linux or online version of a desired Windows program, there might be a Linux counterpart of comparable quality. LibreOffice was one of the best-known examples, offering what many users considered an acceptable, and more affordable, alternative to Microsoft Office. Many Linux programs would be adequate and in some cases superior to their Windows counterparts. Others would not. Searches (1 & 2) led to lists of the best Linux programs, such as Beebom’s list of the 30 Best Linux Apps for 2015, as well as lists by the Linux Alternative Project and Dedoimedo. It would typically be much easier to install and try such programs in Linux than in Windows, thanks to package management software that would permit the user to select, download, install, and uninstall a number of programs at once.

Where no acceptable Linux solution could be found, it was sometimes possible to run the desired Windows program within Linux. The classic tool for this purpose was Wine, perhaps assisted by an optional frontend called PlayOnLinux. Another tool, CrossOver ($59.95, after a two-week free trial), offered an updated and refined version of Wine. Regarding such tools, Ubuntu warned, “Wine is not a perfect replacement for Windows; in fact, it is rather limited. Most Windows programs will not work under Wine; some will be buggy, and a few will work well.” Wine’s FAQs said,

Unlike the biweekly Wine releases, CrossOver releases are rigorously tested for compatibility with CodeWeavers’ supported applications in order to prevent “regressions”. CodeWeavers employs a large proportion of the Wine developers and provides a great deal of leadership for the project. All improvements to Wine eventually work their way into CrossOver.

Both Wine and CrossOver offered databases indicating which Windows programs were supported. In addition, one commentator suggested that forum discussions sometimes contained solutions even when the databases reported failure. Regarding the Wine database, Wine’s FAQs advised that Windows programs rated Platinum, Gold, or Silver were “probably okay” for use in Linux, but those rated Bronze or Garbage probably were not. As a sampling of Wine’s database, a search for Excel revealed various outcomes, depending on the version in question. For instance, Microsoft Excel 2010 had Gold ratings from two Linux Mint users in 2015-2016, and Silver and Bronze ratings from two Ubuntu users in 2013 and 2015. A search for Photoshop gave ratings as high as Platinum for some versions, while a search for Acrobat yielded nothing above Silver for any version released within living memory. Games dominated Wine’s top-10 Platinum and Gold lists. Meanwhile, CrossOver’s database allowed Mac and Linux users to enter ratings of up to five stars. In that database, Linux user ratings averaged four stars for Excel 2003, 2007, and 2010; some versions of Photoshop (e.g., 7.0, CS4) had five stars; but Acrobat versions (aside from 8) were mostly in the one- to three-star range. From this brief look, it seemed that the decision to upgrade from PlayOnLinux to CrossOver might reasonably be driven by the need for a potentially improved Windows emulation for specific programs and/or by the reportedly improved CrossOver user interface (see ZDNet).

Where Wine and CrossOver were unable to install and run a Windows program in Linux, there was the option of trying again with that program’s portable version. There could be different ways of finding a portable version in Windows. First, many Windows programs were available in both installed and portable versions. But when a portable version was not offered, Windows tools such as Cameyo, Evalaze, ThinApp, PortableApps, Ceedo, JauntePE, jPort, LANDesk, BoxedAppEnigma Virtual Box, Turbo (formerly Xenocode or Spoon) Studio, winPenPack, Lupo PenSuite, LiberKey and others could create one. Prices for such tools ranged from free to $2,000+. It appeared that such tools would typically attempt to take a snapshot of the system before and after installation of the desired program, and would then package the differences between the two, so as to help a computer function as if the installation process had taken place. A How-To Geek webpage illustrated the process using Cameyo; a TechTarget page did the same with ThinApp. It did appear, from such illustrations, that at least with the expensive tools, it would be possible to virtualize even some expensive, high-profile programs. Sources suggested that this process would work best on a clean machine, presumably meaning a new Windows installation. Another Geek webpage described other techniques for making portable programs. None of these tools or techniques would work with every Windows program — for example, virtualization might be infeasible for programs that installed or required certain kinds of drivers.

Running Windows Inside or Along With Linux

Instead of looking, one at a time, for Linux alternatives to favored Windows programs, there were techniques for running the entire Windows operating system inside Linux, or in connection with Linux. One such technique would be to run a full installation of Windows 7 in a virtual machine (VM) on Linux. On any operating system, a VM was simply a workspace in which the desired operating system would function as if it were running alone on its own computer. So in one approach, the user would install Linux on a computer; would use VM software to create a VM on that Linux system; and would then install Windows 7 within that VM, typically using the Windows installation DVD.

(While the focus here was on Windows 7, it was also possible to run Windows 10 in a VM on Linux. Doing so could provide access to future Windows 10 applications and capabilities not available in Windows 7, without making one’s entire system dependent upon Microsoft. Note that it was also possible to run various Linux distributions — and also, apparently, Mac OS X — in a Linux VM.)

The VM scenario had potential complexities. First, it would reportedly require a Windows 7 license. For better and for worse, the user who got used to working on Windows within a VM on Linux might not be very motivated to find Linux alternatives to favored Windows programs. That would not necessarily be bad; it might just mean that the user would not become very familiar with, or involved in, the details of the Linux operating system. Another concern, of potentially greater import, had to do with performance. Personal experience with Windows XP in a VM suggested that Windows 7 might run more stably, but also more slowly, than a native Windows 7 installation on bare metal — but there also seemed to be indications that even some gamers were satisfied with modern VM performance.

There were other ways of getting Windows 7 inside a VM. Instead of installing from a Microsoft installation DVD, it might be possible to restore Windows from an image or other backup of the user’s Windows 7 (or 10) drive C, created with Acronis or perhaps with other imaging or backup software. For those whose Windows 7 (or 10) systems had become unstable, it could be helpful to restore an older, stabler Windows image to the VM. It was not clear how a Windows 7 VM would fare, if Microsoft did deliver dysfunctional updates; it was also not clear to what extent a VM might require such updates in order to keep running well.

In addition to the Linux host operating system and the Windows guest operating system, a VM setup would require VM software to enable the guest to run on the host. There were many virtualization programs (Wikipedia). A search led to various (1 2 3 4 5 6) sources suggesting that the best and/or most popular virtualizers for Linux were VirtualBox, VMware, and perhaps KVM, with VirtualBox seeming dominant. Others included QEMU, Xen, and coLinux. There appeared to be some compatibility among such programs. As a potentially absurd illustration, it was at least conceivable that one could use the free VMware Player to create a portable Windows 7 VM that one could then convert to VirtualBox format and run from a USB drive using Portable VirtualBox, while running a Windows 10 VM on installed VirtualBox at the same time.

For some purposes, the user might not have to create or restore his/her own manually built Windows 7 installation in the VM. There was such a thing as a prepackaged Windows 7 VM, ready to run within a program like VirtualBox. At this writing, it appeared possible to download a complete free VM (choosing from Windows XP, 7, 8, or 10) from Microsoft, and run it within the user’s preferred virtualization program on the user’s Windows, Mac, or Linux system. For Linux systems, the Microsoft page offered somewhat confusing options of downloading various versions of the Internet Explorer (IE) or Edge browsers with accompanying copies of the relevant operating system, for use on either VirtualBox or Vagrant virtualization programs. For instance, what the page described as a Linux-based download of IE11 on Windows 7 was a 3.5GB file, and the sole Windows 10 option (MSEdge on Win10) was 5GB. At the time of downloading, the Microsoft page offered installation instructions, and warned that the VM would expire after 90 days, but advised that setting a snapshot upon installation would permit rollback to the starting point as the 90-day period neared its end. It was not clear whether such VMs would then be able to run indefinitely with those 90-day rollbacks, or whether some ultimate time limitation was built into them. (Note also the option of downloading and resetting a 90-day free evaluation copy of Windows 10 Enterprise edition (above).)

If a VM did not work or was not ideal, there were other ways to gain access to Windows programs while using a Linux computer. How-To Geek suggested trying a remote desktop. The idea here was that the user would use Linux to take control of a networked Windows machine, and would then use the software installed on the Windows machine. (Depending on physical location, of course, the user might simply turn to the Windows machine and perform the desired task directly.) To use the remote desktop approach, another Geek article explained that Windows XP, Vista, and 7 had a Remote Desktop option: run SystemPropertiesAdvanced.exe > Remote tab > Allow connections only from computers running Remote Desktop with Network Level Authentication > optionally, Select Users. This reportedly did not work where the remote system was running a Starter or Home version of Windows. Other utilities reportedly offered similar functionality. From a Linux machine, the rdesktop package would apparently permit access to a Windows machine that was configured in the manner just described. Presumably computing speed would be limited by the speed of the network connection (probably faster via direct cable than through a router) and by the resources of the Windows machine. This scenario might be helpful where the user found that one computer was capable of running the desired version of Windows but another was not: the latter could become a Linux machine, accessing the Windows machine as needed.

There was one other way to do Windows computing in parallel with a Linux system: dual- or multi-booting. A machine set up in this way would have two or more operating systems installed (see e.g., Ubuntu’s guide). At bootup, the user would indicate which OS s/he wished to use. It was feasible to arrange things so that the user’s data files would be visible to all installed OSs. So, for example, the user could boot into Windows to perform a task with Windows-specific software; save his/her work; and then reboot into Linux and proceed to use those Windows-created data files. (It was also apparently possible to set up a system to dual boot Windows 7 and 10.) Needless to say, dual-booting could be cumbersome; nonetheless, it could be all that some users would need. Among other things, such an arrangement could facilitate a transition to Linux without disrupting a Windows installation.

Instead of installing Linux onto the computer’s solid-state (SSD) or hard drive (HDD), another multiboot option would be to use a standard or customized Linux live CD. A live CD was an installation CD (or DVD) that had the ability to run, not merely to install. For example, Microsoft could have offered Windows 7 on a live DVD: doing so would have given users the option of installing Windows from the DVD onto the hard drive, as usual, or running Windows from the DVD, without any installation. (It was reportedly possible for users to create their own live Windows CD or USB drive, though it appeared that was easier with Windows 8 and 10 than 7 (see Windows To Go).)

Many Linux distributions offered live CDs. A customized Linux live CD would capture the ways in which the user had modified some version of Linux to suit his/her needs. Needless to say, CDs and DVDs would be able to store only a small operating system configuration. But it appeared that a live Blu-Ray disc would be possible, and it was also possible to boot from large-capacity USB drives and SDHC cards — or, for improved responsiveness, from an external SSD or HDD or from a RAM disk.


Part One of this post discussed Microsoft’s offer of a free upgrade from Windows 7 or 8 to Windows 10. That discussion focused especially on the question of what kind of company Microsoft was, or was becoming, and what its corporate culture might imply for PC users who desired long-term computing stability.

Microsoft was clearly trying to make Windows 10 an operating system that would work with everything from the desktop PC to the Internet of Things. This choice appeared to make Windows less specifically focused on the ideal desktop or laptop computing experience, and that seemed to explain much of the hesitation, among Windows 7 users, to upgrade to Windows 10. Another remarkable drawback of Windows 10, in the eyes of some users, was the open admission that Microsoft would be forcing millions of ordinary Windows users to function as guinea pigs — to use updates that they could not resist, so that Microsoft could find out what was wrong with those updates, and make corrections, before issuing the improved updates to the business customers who provided the bulk of Microsoft’s revenues.

In its desire to drag a billion devices into Windows 10, so as to build up a solid test-bed of guinea pigs, Microsoft had engaged in a series of aggressive maneuvers that seemed to be alienating numerous Windows users. In the perceptions of some users whose Windows 7 installations had been very stable until those aggressive maneuvers began, it appeared that Microsoft might be promoting the switch to Windows 10 by actively sabotaging Windows 7. Be that as it may, the general sense, arising from the discussion in Part One, was that Microsoft had repeatedly demonstrated a willingness to impair Windows operating system stability (in e.g., some aspects of the upgrades to Windows 8 and Windows 10), when doing so would advance Microsoft’s corporate objectives.

Such impressions appeared to be feeding an interest, among some Windows users, in migrating to a different operating system. As a practical measure, it appeared that such users could still obtain their free Windows 10 upgrade (in case they decided to stay with Windows) by making images of their Windows C drives before and after the upgrade and activation, so as to be able to choose between Windows 7 or 10 installations as needed. It seemed that users who wished to continue using Windows 7 or 8 might also want to take certain steps to reduce or end the Microsoft nagging related to the Windows 10 upgrade, and perhaps to delete large Windows 10 upgrade files that might have been downloaded onto their computers without their knowledge.

Beyond those short-term measures, there was a question of whether a user should seriously consider switching to a different operating system. One possibility was Windows 10 Enterprise LTSB, whose purposes and upgrade regime seemed potentially compatible with what some Windows 7 users would like to see. Unfortunately, LTSB did not seem to be intended or affordable for ordinary users. Hence, the discussion turned to non-Windows alternatives. A brief comparison among Windows, Linux, and Mac systems led to the impression that Windows users might find the look and feel of Ubuntu and Mint versions of Linux to be relatively familiar, and might appreciate the opportunities in Linux for customizing how their systems looked and worked. Linux also enjoyed a reputation for significantly lower cost and long-term stability than Windows.

On the question of which programs a Windows user might desire in Linux, it appeared that some Windows programs were available in Linux versions, some Linux programs offered functionality comparable or superior to that of their Windows counterparts, and some Windows programs could be made to run on Linux with the aid of Wine or CrossOver. Where those possibilities failed — or for a solution that would be in some ways more comprehensive and perhaps simpler — the user could run Windows 7, 8, or 10 (or Mac) operating systems within virtual machines in Linux; use remote desktop software to control a networked Windows machine from within Linux, or simply run Windows and Linux machines side by side; or use dual- or multi-booting among Linux, Mac, and/or Windows 7, 8, or 10, as needed, to take advantage of software or hardware capabilities not available on other operating systems. Based on positive prior experience with Windows XP in a VMware VM on Linux, it seemed that a Windows 7 installation within a Linux virtual machine might be protected from future instabilities that might befall native Windows 7 installations.

At this writing, it appeared possible that Microsoft would improve its handling of the transition to Windows 10, would acquire greater sensitivity to the computing needs of its Windows 7 users, or in other ways would enable Windows 7 users either to stay with Windows 7 or to move to Windows 10 without undue disruption of their day-to-day computing needs. On the other hand, it was also possible that Microsoft would continue in its present vein, to the point that a switch to another operating system would seem increasingly advisable to many users. It was possible that Windows 7 would cease to function stably on some computers. It also appeared to be settled that Windows 7 would not run on the latest Intel CPUs. It seemed, in short, that the individual user would have to choose the point, if any, at which s/he should use Wubi, or should use one of the foregoing virtual or parallel computing options, to learn more about Linux and, perhaps, to move toward a transition to Linux.

This entry was posted in Uncategorized and tagged , , , , , , , , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s