The Windows 10 future: Imagine a boot stamping on an upgrade treadmill forever The advent of Windows-as-a-service means that businesses adopting Windows 10 will need to ensure they can monitor their software portfolio for compatibility with Microsoft's latest updates. So says Annette Jump, a research director at Gartner who today addressed the firm's Infrastructure Operations & Data Centre Summit in …
We've seen the start of the Android Desktop releases, now there's an actual PC out. This sounds like the timeframe needed to ramp them up and refine them will match exactly the one needed for Windows 10 to die. Imagine what that'll be like when Google start pushing their own...
Mark my words. There is a piece of software coming which will send Redmond into a bigger panic than Camerons aids when he mentions he's thinking of helping rural farmers by holding a photo shoot in a pig sty. Things are not mature enough for it to be needed yet, but when they are, this is what everyone will be talking about. You heard it here first, folks. It's the Windows -> Android desktop migration suite.
I last used it to download Linux Mint so I appreciate the sentiments. However now that Android is as bad at security/malware as older Windows versions were it is coming nowhere near any desktop machine I have any control over.
ChromeOS maybe, as that's very secure. Android? You'd have to be a lunatic to run that as a desktop OS unless Google seriously up their security gamr..
Android really isn't suitable - each OS update requires input from various OEMs. That's why Google developed ChromeOS.
However, many of the people who might move to ChromeOS - i.e those not dependant on Windows applications - may well have already moved to some mainstream flavour of desktop Linux.
If Google do a desktop OS it will almost certainly be based on the best of the Linux distros out there now but be Android compatible, so can still be called Android desktop. Forget the Android you see on todays phones, just think some ultra hardened Mint or Ubuntu but with the ability to run Android apps almost as a bonus. Microsoft should be scared. Very scared.
Every day more and more legs are kicked out from under the once mighty microsoft. They must despair watching their empire crumble, even though they keep very quiet about it, for example the way it's trying to wrap up its comedy antics in the mobile space.
Who'd have thought the desktop OS market would open up again, and all entirely of their own doing.
Right on cue.
As more and more pieces are put in place here, shuffled around there, the end goal becomes clearer every day.
...will spend a lot more time checking for compatibility of their product with current versions of Windows than they will actually writing "productive" software.
Will there be advance notification to such vendors of the details of projected Windows Builds? I doubt it.
Why are Windows 10 updates any different from earlier Windows? Updates don't change the Windows APIs that applications are programmed to.
If they introduced a major revision to Windows that affected application compatibility, they would provide an upgrade tool for checking compatibility, and offer to run older application in compatibility mode.
The difference with the windows 10 update model, when compared to previous update models is that is completely different.
Microsoft has moved to a constantly evolving version of Windows, which brings in the concept of rings. Most consumers are on a ring that delivers updates on a similar frequency to previous versions of Windows.
Some businesses will be on a slower ring that delivers the same updates. Most businesses who want total control of their organisations IT infrastructure will be using Windows Enterprise, which let's you choose when you want to update.
To add further confusion, some windows 10 users might also be part of the windows insider program, which puts them on the Fast Ring, this doesn't just deliver updates faster, but also provides things in beta, like bash on Windows
Why are Windows 10 updates any different from earlier Windows? Updates don't change the Windows APIs that applications are programmed to.
Win10 updates ARE different.
Updates for 2K, XP/2k3, Vista, 7 and 8/8.1 were separated into critical ie. security and bug fixes, and recommended functionality changes. However, neither of these sets of updates introduced changes in the API's that were incompatible with the pre-existing version. Hence the net result was for example XP gained the capability to interface to secure WiFi networks and use SATA disks, without causing pre-existing applications to fail - yes they might not be able to 'see' the new functionality.
When MS wanted to introduce more fundamental updates that would change API's etc., they issued a new version of .NET or Windows; which users were largely free to install as and when it suited them.
Although MS have made various commitments over disruptive change to Windows 10, the frequency of updating and the total lack of any distinctive user prompt - such as GWX, means in practical terms there is little real distinction between all these style of updates.
Given the update frequency for Win10 and Ms's general stance since the introduction of Win10, I don't expect MS to produce a compatibility tool. My expectation is that they will expect third-parties to have done the relevant testing etc. hence if an application breaks on a user system as a result of a Win10 update, it will be deemed to have been the user's fault as they hadn't downloaded the relevant updates...
No, but I could site on a 5 year LTS version of Linux for the best part of that time.
But as you say, as soon as its "as a service" you basically have to jump to their tune: OS change breaks some bespoke application? Tough shit, pay them to fix it. What, that updated version is not compatible with your archive of valuable data? Tough shit. Office 365 or Google docs has played "hide the feature" again? Tough shit, retrain your staff or stop using it.
Or 10 year if you went with RedHat
Yes I know RH is just as evil as MS in the minds of some here but if you want stability for longer than 5 years then there is a choice.
With MS now injecting App Ads into your start menu, they really don't want people to use their software in the long term (on the desktop). Well, they won't if they keep up these repeating footgun incidents.
They might have reached the tipping point now and the costs of moving to a non Windows platform may now work out less than keeping on it. I can see this 'rapid release' cycle getting shorter and shorter and IT Depts all over the place getting stuck in frantic evaluation and testing cycles every month. The Admins will hark back to the old days and Patch Tuesdays with fond memories.
"Or are you just going to sit on the same version of Linux forever?"
I typically don't upgrade my BSD machines [other than patches] to bleeding edge. Why bother if it's working? patch the infrequent vulnerabilities. no need for 'bleeding edge' if what you have already works. It's why I'm sticking with Windows 7 for my windows boxen.
As for Linux, I typically stick with one version of THAT as well. It's more stable, particularly for software development. If I upgraded a particular ubuntu build machine (a virtualbox VM) that has patched compilers [patched by me] for a particular CPU, I'd have to re-do the patches. better to leave it 'as-is', because it's *STABLE*. So what if it's 3 years old.
'Bleeding edge' is *OVER*-*RATED*
>"those still struggling to get rid of IE & ActiveX crap are in for a massive re-wire effort either way."
>>Which gives them the opportunity and reason to make a long-term decision.
I get the impression that whatever one uses to replace IE5 and ActiveX is platform-agnostic. That is, people having learnt their lesson about getting stuck in the mud before will not make the same mistake again, and make the decision to keep their options open in future. I'm no expert, but it seems that if even productivity software such as 3D CAD can now be run through a web browser, the actual OS of the desktop computer (er, terminal?) doesn't matter so much any more for many tasks, so long as it's secure and reliable.
You clearly have no idea of the resource a true 3D CAD application needs. A demo of a browser running a CAD application doesn't make it a viable solution for professionals. But keep on believing in fairies and unicorns...
BTW: high end CAD application do run on some Unix workstations, but there's nothing for Linux. There are also many other specialized applications for designing specific systems, and again, you may not find them running on Linux.
Often the GPL and the lack of proprietary drivers fully exploiting the hardware are a roadblock for wider Linux adoption. The ZFS issue is just an example of how difficult could be to develop kernel modules without giving IP away.
"The ZFS issue is just an example of how difficult could be to develop kernel modules without giving IP away."
That shows a complete misunderstanding of the situation. Firstly virtually no "applications" need any kernel modules, typically that is for special hardware and things like file systems. Secondly you can develop a kernel module and make it available as a binary blob to be added to someone's Linux system if you want - after all that is what Nvidia, etc, do for graphics drivers. The current argument is about a distributing the GPL Linux kernel with a pre-compiled non-GPL driver and if that makes it "distributing a derivative" of the kernel (which seems a bit bizarre argument).
The lack of specialist applications for anything other than Windows is simply a historical artefact of 90+% of desktop computers being Windows based, why would you bother with the other 10%? However, if a lot of folk move off Windows due to this, or other reasons, then software developers may start to see the value in using cross-platform tools (like Qt and similar) so they are not tied to MS uncertain future roadmap.
Or just run stuff in a Win7 VM without email/web/external Internet access and forget about the future patching (or lack of) for the OS.
How does all this time and effort and expense compare to migrating to Linux?
Apples and moon rocks in many cases: many companies are, more or less, happily wedded to Windows stacks.
In any case the main migration that Microsoft is worried about is desktop to mobile (IOS or Android).
That just depends on what software you're using - and that's what Microsoft is betting on. The average enterprise environment is a bit more complex than a small office setup.
You may be using applications for which there are no _professional_ Linux alternatives (albeit there could be OSX ones - and that can be another migration path), or there are no Linux version at all (i.e. specific and custom software, maybe running expensive devices). WINE may be too risky to be used in many situations.
There could be also the whole management issue, you may like it or not, but Windows desktop and users are easier to manage using Active Directory and related tools. Sure, Puppet and the like can help, but you have to factor in also the costs of migrating all your IT skills to different tools, and the risks of more errors while getting proficient involved.
Maybe one day Linux distro will agree on a common AD competitor (with decent tools) and will make Linux even a more compelling alternative even on desktops.
"You may be using applications for which there are no _professional_ Linux alternatives "
That could certainly apply if the application is abandoned already, in which case some migration strategy is needed anyway.
More relevantly, what do you think is going through the minds of _professional_ Windows-dependent app developers and vendors at the moment? Do you not think that many of them are quietly evaluating their next generation options, prompted by MS making it increasingly clear that Windows is becoming a cloud service, either Windows 10 Enterprise on a corporate cloud, or Windows 10 everything else on an Internet cloud?
Cloud doesn't suit everyone. Not every user, not every admin, not every developer.
Follow the money.
You could be surprised how many users care only if their needed applications are available and run, and care little about the underlying OS.
Windows 10 may be an issue for many of us who don't like slurping and may have to support it in the harsher environment MS has decided to create, but users won't care until, and only until, their applications have issues.
If there were the same application on Linux, you could easily move them. The problem is Linux lacks too many professional applications, for several and different reasons. When clones exists, some are good, but often not good enough. There are some de facto standard applications you can't really work without unless you're looking for troubles. Often, there's no replacement because they are very specific and/or custom made applications.
I've seen many Linux desktop replaced by OSX ones because of applications availability. You can still run most *nix software, and also all the OSX applications. I've seen a lot of developers also getting OSX machines when they could.
Even if they are abandoned applications, they may still run on a recent Windows. Replacing them with Linux ones? While you can find a lot of "Linux" developers if you can replace them with web ones, there are far less that can develop desktop applications. Linux never got friendly IDEs and frameworks to develop applications quickly, and there are far less developers available, making them more expensive also.
Windows 10 may be an issue for many of us who don't like slurping and may have to support it in the harsher environment MS has decided to create, but users won't care until, and only until, their applications have issues.
But at the heart of this is MS's fundamental problem and obsession, namely maintaining customer awareness of the Windows brand. Hence why, just like an attention seeking child, they feel the need to find new ways of thrusting it into users faces and drawing attention to it.Even though, as you point out, users are more interested in being able to access and share documents across a wide variety of devices and OS's.
Unfortunately, what MS haven't realised is with a subscription model it doesn't matter if users forget about the OS, just so long as they keep paying the subscription. In fact with a subscription model, I suggest it is more important that change happens smoothly and so users do forget, otherwise the service will gain a reputation for being unreliable and users will switch...
I can only speak for myself, but my current Win7Pro64 machine will be the last one I own with Windows on it. Between MSHQ's attempts to force Win10 down my throat, Win10's total lack of ability for me to choose which updates to apply if any, the fact that those updates have a VERY strong probability of rendering my screen reader (& by extension my computer) inoperable, constantly ignoring my user settings in order to "entice me to upgrade" (What part of "No Means No!" do they not comprehend?), I'm sick & fekking tired of the MS treadmill.
I've been researching my next computer, trying to decide between Apple & Linux. I need the Screen Reader Environment (SRE) in order to use the computer at all, & Apple has the best SRE by far hands down. Unfortunately they'll charge me more for a 4th gen Dual Core 3GHz with 8Gigs of RAM & a 250GB Sata-3 SSD than I'm entirely comfortable with paying. For what they want I can save over $200 & buy a 6th gen i7 4GHz with 32Gigs of RAM & a 250Gb M.2 SSD from System76 instead. Apple may have the best SRE, but I can't see paying MORE for mere 4th gen hardware, especially not when I can use the savings to buy a Support Contract from S76 *on top* of the 2 year warranty.
And that's what it's boiled down to, Apple or Linux, older hardware but a top notch SRE versus brand new hardware & a hopefully working SRE.
MS has done a great job of pushing their new OS, they've pushed me so hard I'm jumping off the Windows treadmill & getting a Linux machine instead.
Well the evidence from some years back is that the actual cost of a desktop migration from Win X to Win Y is very similar to Win X to Linux. However, the areas where costs are incurred are user training and (initial) support. Although this does assume both the equivalence of Windows and Linux applications (eg. MS Office == LibreOffice) and that all business applications support a Linux client.
However, this analysis of MS's Win10 update policy announcement, if correct, significantly increases the level of on-going build support a business is going to have to perform. Which will start to make the use of a LTS Linux desktop build much more attractive, particularly as I suspect few businesses will look kindly on an IT department asking for an increased budget with no commensurate bottom line benefit.
"Got a pop-up yesterday telling me when my Windows 10 upgrade was scheduled. Microsoft need to be hit with the world's biggest anti-trust suit."
I'm waiting for a 'The Register' article on this. Saw one on Tom's Hardware, e-mailed URL to Reg staff yesterday...
http://www.tomshardware.com/news/windows-10-auto-schedules-updates,31802.html
El Reg, this one's pretty heinous. As bad as the malware-looking "upgrade now" or "upgrade later"
I didn't get that pop-up. Looks like GWX control panel is working.
Win 10 may be better on various fronts, but I resent the way they are trying to force feed us the OS, as much as I resent the data slurping and forced upgrades embedded in it. I have several data capture applications that work under Win 7 and 8.X. When capturing a long sequence of data (like the 176 GB from the Mercury transit), I do not want this interrupted by some upgrade MS foists upon me. I do tend to keep my system updated, but I also do want some degree of control.
Came home to something similar yesterday, except that one was "Install now" or "Install tonight". Nowhere was an option given to not install, although closing the window without clicking either button will do it. This crap is unforgivable.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
