Windows 7: Five things Microsoft must do

Windows 7: Five things Microsoft must do

Summary: A lot of businesses have already decided to boycott Windows Vista and many consumers would like to skip it, too. In order to avoid the same mistakes in the next version of the operating system, there are five things Microsoft needs to do with Windows 7.

SHARE:

We're going to be hearing a lot about Windows 7 over the next two weeks as Microsoft convenes its Professional Developers Conference (PDC) and the Windows Hardware Engineering Conference (WinHEC) -- both in Los Angeles.

I recently discussed the five reasons why Windows Vista failed, and although Microsoft may continue to defend Vista as Steve Ballmer did at the Gartner Symposium on October 16, make no mistake that the company has already moved past Vista -- psychologically and strategically.

Ballmer and company know that Vista has been rejected by businesses and that it has dreadful reputation among the general public. They know they screwed up. And Microsoft is nothing if not responsive to customers and public opinion. In fact, I've argued before that the company is actually hyper-responsive and that's why they've crammed way too many features and way too much code into both Windows and Microsoft Office.

So how can Microsoft recover from the Vista debacle and breath life back into Windows? If nothing else, here are five things the company needs to do to start.

5. Speed it up significantly

One of the worst qualities of Windows Vista is that it is almost always slower than its predecessor Windows XP when running on the same hardware. And most of that performance drag is not simply due to the fancy graphics. Even if you turn off the Aero interface, Vista is still usually slower.

The problem is that the underlying Windows code is way too bloated -- over 50 million lines of code in Vista -- and even today's ultra-fast multi-core processors can't overcome that. The Windows development team has to find a way to streamline Windows 7 so that it's faster and more stable than both XP and Vista, whether it's running on netbooks and nettops with the Atom processor and only 1 GB of RAM or tomorrow's 8-core machines with 10 GB of RAM.

That's an extremely difficult task, but no one said this was going to be simple or easy. One way to start is by turning Windows into just the core OS and further modulizing it by making a lot of the other software such as the Media Center, Tablet PC, and Admin Tools available as downloadable add-ons.

4. Avoid compatibility problems

In the process of streamlining Windows 7, the developers can't sacrifice software compatibility. One of the things that has killed Vista is that Microsoft spent so much effort trying making it more secure with User Access Control (UAC) that it broke a lot of software in the process.

You can argue that a lot of the stuff that broke in Windows Vista was poorly programmed to begin with and deserved to break so that it could be rewritten more securely. The problem is that not much of the software has been rewritten and the UAC approach has not worked because users get so many dialog boxes that they just blindly click OK until all of them go away. A better approach is needed -- one that balances security and compatibility.

The other compatibility issue that Windows 7 has to juggle is the 32-bit vs. 64-bit split. While most modern processors are 64-bit, most of the software and device drivers are still written in 32-bit code. I've seen a number of PCs with 64-bit CPUs that have 32-bit Windows installed simply because it has better compatibility. I've also seen and heard about a number of business systems that have 64-bit Windows Vista installed, but are running into significant software and/or driver incompatibility problems.

Microsoft, Intel, and AMD need to lead the charge to get software vendors on-board with 64-bit before Windows 7 is officially released.

3. Undercut OS X on price

Mac sales have been growing much faster than the overall PC market and Mac OS X has continued to nibble away at Windows' massive market share over the past two years. However, Apple showed the same chink in its armor that has long plagued it when it recently announced its new line of laptops and the cheapest one was priced at $999. The message being sent is that Apple wants to be a premium computer brand with high margins and has very little interest in selling low-margin, high-volume machines.

Over the next two to three years the lion's share of the growth in computer sales is very likely going to be in the sub-$500 netbook and nettop market. These machines are essentially just glorified Web browsers in a diminuitive hardware package. The OS doesn't matter much. As a result, Linux is a major threat to become the OS powering a lot of these machines, because of its minimal price.

However, with Apple relegating itself to the high end of the market and most users still not very comfortable with Linux, Microsoft has the opportunity to swoop in and deliver a Windows 7 that is fast and cheap and can run very well on these little machines, while also scaling all the way up to the fastest workstations. A lot of users and businesses would probably gravitate toward the idea of a common OS experience (and one that most users already know) in Windows, especially if the price is comparable between Linux and Windows machines.

The key here is making Windows very inexpensive and very scalable while preparing to sell it in larger volumes than ever before on the cheap machines that are going to flood the market over the next couple years.

2. Sell only one version

There were primarily two editions of Windows XP: Home and Professional. With Windows Vista, that doubled to four primary editions: Home Basic, Home Premium, Business, and Ultimate. It's time to simplify and go back to just one version of Windows with one price.

This is a case of Microsoft just getting out of the way and letting Windows be Windows. Having just one edition of the client OS will make Windows 7 easier to understand, easier to purchase, and easier to support (for both Microsoft and IT departments).

Of course, the one version of Windows 7 needs to be cheaper than Mac OS X ($99), easier to use than Linux, and easier to set up and get started than any of the recent versions of Windows.

1.  Make it the last shrink-wrapped OS

The old way of building and packaging operating systems in shrink-wrapped boxes that are released every few years is just not fast enough or nimble enough to meet the demands to today's Internet-driven computing environment. It's also counterproductive for an OS maker because you end up competing against yourself the way XP and Vista are now competing against each other.

There's only one Windows, and it has merely evolved over time. That's the message Microsoft needs to drive home by making Windows 7 the last shrink-wrapped version of the OS. From here on out, Microsoft should simply make Windows a constantly evolving platform with new features and functionality enhancements added several times a year through Windows Update.

The business model would be to turn this into a subscription product, albeit a very inexpensive one. As long as you have a current Windows subscription then you can continue to download new features, patches, and updates. If your subscription lapses then Windows still works but you can no longer download the new stuff, or any add-ons, and you can only download highly critical security patches.

For enterprises that are currently using Software Assurance, they are already buying Windows as part of a subscription so there would be no change in the business model for them.  For consumers and small businesses who aren't part of Software Assurance and typically buy Windows from OEMs such as Dell, Toshiba,  and Hewlett-Packard, the Windows license that comes with their PC could last for three years and then it's up to the buyer to pay something like $30-$40/year to renew. For those who want to build their own system, a full version of the OS could cost something like $50-$75 for the first year.

Bottom line

Windows 7 needs to be fast, inexpensive, and widely compatible. Microsoft also needs to change the development and business models to make Windows one continually evolving OS.

Let's face it, the OS is not as flashy as it once was. It's also not nearly as relevant as it was a decade ago. The Web browser is gradually usurping its position as the most important application platform, as has long been predicted.

Because of that, Windows is at a crossroads where it could begin losing large chunks of market share to competitors that are better prepared to operate in this new reality, or it can greatly simplify its OS while turning into more of a background utility that makes good money off of a low-margin, high-volume business.

If it can pull that off and clearly communicate to businesses and consumers that Windows 7 is the start of a new approach to Windows then Windows 7 could be a watershed release. If Microsoft simply releases a mild revision to Vista and maintains the same development and business models, then Windows could become more vulnerable to its competitors than it's been in almost two decades.

Topics: Windows, Microsoft, Operating Systems, Software

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

101 comments
Log in or register to join the discussion
  • Should Windows go incremental?

    Would you welcome more incremental updates to the OS rather than a big release every few years? Why or why not?
    JMHtech
  • RE: Windows 7: Five things Microsoft needs to do

    Great article.

    I don't think the incremental approach would work well
    for most consumers because the type of updates that
    would potentially need to be rolled out between major
    versions of Windows would take too long.

    We went from CD's to DVD's with Windows Vista.

    While it can be argued that the code is bloated, it
    would still take way to long to roll out major updates
    to an OS via updates.

    They are too large and you can not assume everyone has
    broadband.

    Oh, and that would be fun when they fail halfway...

    (nightmare)..

    Onuora Amobi
    CEO, Nnigma inc.
    www.amobi.com
    www.windows-vista-update.com
    Ammalgam
  • Interesting

    Your approach related to the signficant evolution of Windows business model is interesting but i don't think that this could happen with Windows 7 which will be more or less a significantly,hopefully, improved version of Vista.

    However, Microsoft should consider such approach with their next major O.S which will hopefully use virtualization for any kind of backward compatibility.

    Moreover, i don't think that web browsers will be able to replace true O.S for virtually everything anytime soon. For once because browsers are too limited and second because they will need to become much more O.S like to be effectively able to replace them.
    At this point one could argue that what O.S need to do is to acquire some Web browser capabilities ,such as be able to run Web applications by completely bypassing web browsers, to keep their relevance.

    I personally think that the right approach is the one of Microsoft with the software+services concept thus the user can stay productive even if it is offline.

    And finally i don't know if it is a good idea or even it is even possible for Microsoft to release only one version of Windows because of entities such as E.U.
    E.U is the reason of the release of the absurd N editions whom know seem to want.
    One could wonder why EU don't force Apple to release a Mac OS X without browser or any kind of application which could represent an unfair advantage compared to competitors too.
    timiteh
  • Will, since your last article...

    ...was lambasted because of it's somewhat dodgy supporting evidence - in particular where you accused Vista of being slow in comparison to XP by quoting a CNET article released in 2007 when SP1 was pre-Beta - I'm going to take this one with a pinch of salt too.

    But it's a slow day so what the hell..

    "One of the worst qualities of Windows Vista is that it is almost always slower than its predecessor Windows XP when running on the same hardware."

    Blatantly untrue. Here's an article where the journalist actually did some research.

    http://www.extremetech.com/article2/0,2845,2328909,00.asp

    Also, here's the benchmarks proving it:

    http://www.extremetech.com/article2/0,2845,2302500,00.asp

    So, once again, you're shown to be at best off the pace in terms of tech reporting or at worst someone with an agenda. I'll leave it up to your readership to determine which of the two applies.

    "One of the things that has killed Vista is that Microsoft spent so much effort trying making it more secure with User Access Control (UAC) that it broke a lot of software in the process."

    Again untrue. This was largely due to incompatible drivers, something that hasn't been an issue for over six months. If UAC is an issue then it can be switched off although I wouldn't recommend this. I'd also point out that every change of OS has compatibility problems - this was certaily the case when we switched from NT to XP on our corporate platform.

    "3. Undercut OS X on price"

    Since you can't install OSX on anything other than Apple hardware they do. As such your comments about the netbook market are irrelevant because Apple don't operate in that space.

    "There were primarily two editions of Windows XP: Home and Professional. With Windows Vista, that doubled to four primary editions: Home Basic, Home Premium, Business, and Ultimate. It?s time to simplify and go back to just one version of Windows with one price."

    I largely agree with this one. Windows 7 should be modular with the option to install additional features as required on top of a core product.

    That said, I'm unaware of anyone whose actually been confused by Vista's variants.

    "1. Make it the last shrink-wrapped OS"

    I agree with this.

    So some good points and some bad. A step above the last article but you need to drop the whole 'Vista failed' thing especially when your statements aren't backed up by fact.
    Sleeper Service
    • huh? curious

      [i]...was lambasted because of it's somewhat dodgy supporting evidence - in particular where you accused Vista of being slow in comparison to XP by quoting a CNET article released in 2007 when SP1 was pre-Beta - I'm going to take this one with a pinch of salt too.

      But it's a slow day so what the hell..

      "One of the worst qualities of Windows Vista is that it is almost always slower than its predecessor Windows XP when running on the same hardware."

      Blatantly untrue. Here's an article where the journalist actually did some research.

      http://www.extremetech.com/article2/0,2845,2328909,00.asp
      [/i]

      Vista was slower at the time the referenced article was written. Vista was as fast or faster at the time the article you yourself referenced was written.

      So saying Blatantly untrue is negating factual findings from the past.

      Par Example: It's the same like saying Bill Clinton is the US President. It was true at some stage in the past.
      TedKraan
      • But THIS blog was written on 10/27/2008.

        And it claims Vista is slower than XP even though the benchmarks provided show it to be just as fast or slightly faster.

        I am constantly amazed at how these urban legends continue to make the rounds. I've run Vista Ultimate on a P4 1.6GHz system with 1GB of RAM. Its performance was perfectly acceptable for general use.
        ye
        • Because it's not an urban legend

          especially since it seems to be a pretty common complaint from a lot of people around the country. ]:)
          Linux User 147560
          • Just because the clueless masses repeat something doesn't...

            ...make it true. Nor does repeating something ad nauseam make it true. Yet that's just what the clueless think. So they persist.

            Now Sleeper Service provided a link to some benchmarks. Have you had a chance to review them? In those three benchmarks Vista SP1 outperformed or matched XP SP3. Those are solid numbers showing Vista can keep pace with XP. Now I'm sure you can find some numbers showing the opposite. So what does that prove? What we would expect to see. Some OSes are faster ar some tasks whereas others are faster at other tasks.
            ye
          • Actually...

            That's the reason i dodged SP3 for XP.
            TedKraan
          • What is? What point are you trying to make? (nt)

            .
            ye
          • If you want XP just for speed

            dodge SP3.
            TedKraan
          • Clueless masses?

            [url=http://www.vistareview.info/2007/12/10/windows-vista-sp1-just-as-slow/]I am sure this fellow would argue the[/url] difference... that was the first hit from using "complaints of Vista being slow" in Google.

            [url=http://www.pcadvisor.co.uk/forums/index.cfm?action=showthread&threadid=287131&forumid=23]The proof is out there....[/url]

            ]:)

            Edit - Fixed first link... forgot the =
            Linux User 147560
          • First link isn't working. However from the link I can...

            ...see the article is pre-SP1 release:

            ttp://www.vistareview.info/[b]2007/12/10/[/b]windows-vista-sp1-just-as-slow/

            So I repeat:

            Just because the clueless masses repeat something doesn't make it true. Nor does repeating something ad nauseam make it true. Yet that's just what the clueless think. So they persist.
            ye
          • Ubuntu 7.04 to 8.10 Benchmarks: Is Ubuntu Getting Slower?

            http://www.phoronix.com/scan.php?page=article&item=ubuntu_bench_2008&num=1

            The answer appears to be "Yes". Seems Microsoft isn't the only one guilty of releasing an OS that may not perform at the same level as the previous version(s).

            I await your spin.
            ye
          • .....

            Sounds like Canonical needs to do a bit of werk... and since I don't use any of the *buntu's, I don't really care. OpenSuSE on the other hand seems in my experience to be getting faster and sleeker. ]:)

            And if you want a spin... head down to the laundry mat...
            Linux User 147560
          • @Linux User 147560: LOL! Nice cop out!

            Is Ubuntu not Linux?

            [i]OpenSuSE on the other hand [b]seems[/b] in my experience to be getting faster and sleeker.[/i]

            Key word highlighted. Where have we heard something like this before? LOL! You got owned!
            ye
          • Gee the same logic you

            applied can be used to Vista Ultimate and Vista Home. Education time numb nutz

            Linux is the kernel nothing more nothing less.
            The distribution is all the goodies packed on top. So your assertion that Ubuntu is Linux is off base. It uses GNU/Linux kernel and tools but adds it's own package manager, a default desktop and other goodies. This is what differentiates distributions from each other.

            When you finally learn that Linux is not OpenSuSE or Ubuntu or Red Hat but actually the underlying kernel they use as a base then MAYBE you will start to understand Linux.

            So no, you didn't "own" anyone. ]:)
            Linux User 147560
          • For the Linux User

            "Linux is the kernel nothing more nothing less.
            The distribution is all the goodies packed on top. So your assertion that Ubuntu is Linux is off base. It uses GNU/Linux kernel and tools but adds it's own package manager, a default desktop and other goodies. This is what differentiates distributions from each other."

            Without the stuff that is added for the distros... really just makes Linux on par with MS-DOS. People hated using command lines even when it was all they had. Without all the extras, the average user is not going to be interested in taking a step back in time.

            For this reason, users see Linux as the whole package that comes in a distro, and not just the kernel. Lets face it, without the GUI and other tools you might as well crank up Unix without X Windows.

            Until the Linux community wakes up and realizes consumers don't what what the developers think is best for them and start listening to what consumers want, you're not going to gain ground. Apple has struggled to gather 5% of marketshare using the "I know best" method and almost disappeared completely more then once.

            Linux makes great servers. No one can argue that. Their lower marketshare makes them a smaller target. But they suck as a general purpose desktop for the masses. A consumer wants a computer to make them feel smarter, not dumber. Linux just isn't straight forward to 'joe the plumber'.

            So who cares if its Ubuntu or Kbuntu, or Red freaking Hat. The irony is SUSE actually got its start from Microsoft. Before they mae Windows they developed a version of Unix called Xenix. As bad as you might not want to hear it, the evil monopoly helped contribute code the the very OS you use today.

            Feel warm and fuzy inside now? Don't believe me? Look it up.

            Attacking Vista's crappy UAC, too many editions, and other things of these nature are perfectly understandable. However it really gets old hearing how horrible something is simply because it says so on the internet.

            If you believed everything on the net then you wouldn't be here complaining they are wrong all the time.

            I'm sure by now everyone has heard your opinion on why you use Linux. Now let them decide why they do or do not.

            But should you really want to continue and compare Vista with Linux, then please provide actual data to back up your claims. Simply stating your experience with XP and Linux on similar machines is not evidience that Vista sucks.

            It sucks to you. We hear you. Move on.
            ShadowGIATL
          • @ShadowGIATL

            "[I]Linux is the kernel nothing more nothing less.
            The distribution is all the goodies packed on top. So your assertion that Ubuntu is Linux is off base. It uses GNU/Linux kernel and tools but adds it's own package manager, a default desktop and other goodies. This is what differentiates distributions from each other.[/I]"

            [B]?Without the stuff that is added for the distros... really just makes Linux on par with MS-DOS. People hated using command lines even when it was all they had. Without all the extras, the average user is not going to be interested in taking a step back in time.

            For this reason, users see Linux as the whole package that comes in a distro, and not just the kernel. Lets face it, without the GUI and other tools you might as well crank up Unix without X Windows.

            Until the Linux community wakes up and realizes consumers don't what what the developers think is best for them and start listening to what consumers want, you're not going to gain ground. Apple has struggled to gather 5% of marketshare using the "I know best" method and almost disappeared completely more then once.[/B]?

            And yet Apple is growing despite it's controlling attitude. Fiscally they are doing quite well. The rest I can't really comment on.

            ?[B]Linux makes great servers. No one can argue that. Their lower marketshare makes them a smaller target. But they suck as a general purpose desktop for the masses. A consumer wants a computer to make them feel smarter, not dumber. Linux just isn't straight forward to 'joe the plumber'.[/B]?

            But I don't see that as being true. Especially since I took a small portion of time to teach people how to use their PC's, both Linux and Windows. It's amazing how much a little guidance helps. Most Windows users get themselves in trouble and have many of the same issues with Windows as they would with Linux. Especially now with the modern distributions out there.

            ?[B]So who cares if its Ubuntu or Kbuntu, or Red freaking Hat. The irony is SUSE actually got its start from Microsoft. Before they mae Windows they developed a version of Unix called Xenix. As bad as you might not want to hear it, the evil monopoly helped contribute code the the very OS you use today.[/B]?

            Actually no, SuSE did not get it's start from Microsoft. They got their start in Germany in 1992 as a UNIX consulting group. SuSE is an acronym for the German phrase [B]S[/B]oftware- [B]u[/B]nd [B]S[/B]ystem-[B]E[/B]ntwicklung which means software and system development. It was first released in 1994 as SLS/Slackware. In 1996 after integrating with the Jurix distribution S.u.S.E. Linux 4.2 was released.

            As time passed and they matured SuSE started to integrate from Red Hat the RPM package system and /etc/sysconfig. It was in 2003 that Novell bought SuSE and changed the name from SuSE to SuSE Linux. So nowhere in that history does Microsoft give SuSE or OpenSuSE a start.

            ?[B]Feel warm and fuzy inside now? Don't believe me? Look it up.[/B]?

            No I don't believe you because I have been using Linux since 1999 and have followed it's growth and history since then. Also the deal with Microsoft was an agreement not to sue for patent infringement.

            ?[B]Attacking Vista's crappy UAC, too many editions, and other things of these nature are perfectly understandable. However it really gets old hearing how horrible something is simply because it says so on the internet.[/B]?

            And personal experience doesn't count as well? While I have not directly been afflicted with the abomination of Vista, I have seen it first hand from my co-workers and boss. Many of which have either moved to Mac, Linux or upgraded to Windows XP Pro.

            ?[B]If you believed everything on the net then you wouldn't be here complaining they are wrong all the time.

            I'm sure by now everyone has heard your opinion on why you use Linux. Now let them decide why they do or do not.[/B]?

            I don't believe everything on the net but when you see enough of the same information that is from differing parts of the world, differing cultures and value systems, then it lends credibility to it. If you want Linux users to stop then the other camp needs to as well. The lies about Linux are getting old and the rehash of problems long since fixed is rather tedious.

            ?[B]But should you really want to continue and compare Vista with Linux, then please provide actual data to back up your claims. Simply stating your experience with XP and Linux on similar machines is not evidience that Vista sucks.[/B]?

            Are you going to pay for Vista for me to test out? All the various version that are available? And of course the hardware to run it on too.

            ?[B]It sucks to you. We hear you. Move on.[/B]?

            It sucks to a lot of people. ]:)
            Linux User 147560
          • Ye...

            [i]Just because the clueless masses repeat something doesn't make it true. Nor does repeating something ad nauseam make it true. Yet that's just what the clueless think. So they persist. [/i]

            It's worked well for Microsoft though. How many people out there believe that Active Directory is an LDAP?
            914four