Are they dropping legacy apps???
As I understood it, the version number allowed apps to run in compatibility mode if they weren't the latest and greatest. It'll be interesting how this plays out if that was true.
Microsoft surprised us by upping its Windows product numbering from 8.1 straight to 10, and now it appears it's planning to make an even greater leap in the version numbering of the Windows kernel itself. While Windows releases use familiar marketing brands – such as XP, Vista, 7, 8 and now 10 – the underlying operating system …
> Are they dropping legacy apps?
Remember Microsoft's big R&D on emulation? And on Docker for Windows? And dozens of other projects aimed at creating device-independent containers for apps to run in?
Yeah. That's where your legacy app will live, as you stream Office from the cloud. On your cellphone.
"As I understood it, the version number allowed apps to run in compatibility mode if they weren't the latest and greatest. It'll be interesting how this plays out if that was true."
No. The main reason why it's been the same major version number since Vista is to accommodate those awful programmers who solely rely on the major version number to determine whether their apps are compatible or not. For these applications, changing the major version number may cause those programs to insist that your version of Windows isn't compatible, even though there is no actual reason for it to not work.
However, we're in 2014, and those lazy programmers have had nearly 10 years to change their ways, and with VMs and virtual layers within Windows (like the previous poster mentioned) as well as a Windows Store that works differently with regards to versions of Windows, there is no excuse now.
........declare itself to the application/program as version 10 if the app/prog correctly identifies itself as Win 10 compatible. All other apps relying on the version number 6.x will be told a small "porkie" by the os in as much that they will be told the version is a 6.x build. In practice this should mean that any programmes that were compatible with Vista, 7, 8 and 8.1 should not refuse to run on Win 10. At least I imagine that Redmond are keeping their fingers tightly crossed that it is indeed the case!
*Which I do not take as a given. If I am mistaken I will gladly take correction.
For the vast majority of Windows programmers, compatibility is handled by your compiler and runtime library.
Apart from a tiny number of open source app programmers, there are no awful programmers who rely on the major version number to determine whether their apps are compatible or not. And that number is tiny because the vast majority of open source apps run on a unix emulation layer, and have no regard for native os features.
But "9" is a magic number, because if we go back to the 90's, some "awful programmers" relied on the "9" to detect windows 95 and 98 -- particularly in scripts, and often not actually programmers at all, but that was at a time when it was possible for random people to do programming (unless they were using unix).
"those lazy programmers have had nearly 10 years to change their ways"
Unfortunately, those lazy programmers have had nearly 10 years of confirmation that they don't need to change anything and it will continue to work. Older versions of Windows had compatability tables that understood each version of your applications and would deliver a suitable version number through the system calls to trick the badly written apps into thinking they were running on the right type of obsolete system to continue running.
Or Microsoft could try the Apple approach and deliberately break the apps, which would force users to go out and buy new software with each OS release.
"...all part of Microsoft's effort to unite the desktop, server, and mobile versions of its OS around a single, common core."
I would hazard a guess as to that core being Microsoft servers/cloud. I cant help thinking that Microsoft would like Windows to fit the SaaS model. The design look and feel of Windows has slowly evolved such that there has been less and less distinction between local and remote/Internet content. Imho in a few more iterations of Windows there will be no difference at all.
I am sure many large software houses would like to see the end user in a computing environment over which they have little or no control, tied to a subscription.
If you don't already you might want to look into supporting open source alternatives and not just for operating systems.
I use Windows 7 and it's great for the things I use it for.. As it CentOS and Linux Mint, the right tool for the right job, or my preferred tool for the job. Whatever, the point is, I want my data local and I want to choose what my OS does when I am not looking. I want to choose my default programs and utilities I do not want to spend several hours changing settings in order to stop the OS and the web browser and the media player and umpteen other bits of software from phoning home first opportunity they get and henceforth ever after.
I do not want to spend several hours changing settings in order to stop the OS and the web browser and the media player and umpteen other bits of software from phoning home first opportunity they get and henceforth ever after.
If you don't like OS X and Windows there is always Linux.
Then you can do the phoning home for them to get updated drivers, etc.
If they were OK with version 6.4, why change it to 10.0? Seems like they're trying to bring them in sync, making the kernel version useless in the future as it'll probably go to 10.1 when Windows 10.1 is released, to 11.0 when Windows 11 is released and so on.
At least until now the kernel version has been a good indicator of when they are making major changes. For instance, NT 4.0 drivers needed a rewrite for Win2K, but XP was OK with minor tweaks. XP drivers needed a rewrite for Vista, but 7 and 8 were OK with minor tweaks.
If drivers from Windows 8 mostly require only minor tweaks, or will run unchanged, we'll know that the kernel version 10.0 thing is a fantasy, and it is really version 6.4 under the hood.
Is it shady?
Or is MS simply now doing what the rest of us do every day at work?
I mean, when you change the code of a program's that has a version number, don't you increment the version number every single time?
Wouldn't you be sacked if you broke this rule?
MS broke the rule because lazy incompetent outside programmers used version number to determine compatibility. They did this for 10 years.
If you're happy revolving your world around lazy incompetent programmers you should run Linux. According to Linus that what many of his contributors are.
As for me, I'm happy when MS starts to follow industry standards.
They were incrementing the version number. Vista was kernel 6.0, 7 was kernel 6.1, 8 was kernel 6.2, 8.1 was kernel 6.3. Until recently 10 was kernel 6.4, which would indicate that it was a minor rev and not something big that would change the driver interfaces in a major ways and require driver rewrites.
Yes. Sadly it use to be one of the few non-shady things MS could be counted on to maintain.
Major number revisions are to reflect major feature changes. These are supposed to be fully tested. These are expected to be sequential.
Minor number revisions are to reflect improvements and security fixes to the current major revision. These are supposed to be fully tested. These are expected to be sequential.
Point number revisions are to reflect other sort of changes to the code base. These may not be fully tested but are expected to have passed rudimentary tests run by the programming team. They are not expected to be sequential as they are used by the build team to track internal iterations of the program.
You're right that the rest of the industry has adopted Google's Chrome process (and I fault Mozilla for giving them covering fire). That doesn't mean it wasn't shady when Google started it. It doesn't mean it will be any less shady if everyone adopts it.
Something has happened in the OS space, everything has become "minor" regardless of the OS version.... W8 tried "major" changes and failed... W10 is going back to minor changes and yet I still don't see it as a runaway success. I don't have a Mac so I can't compare but if I have understood correctly there are no major changes there either.
If we have reached peak OS, or thereabouts, can't we simply stop all the version numbers and just have kernels where things will be garaunteed to run fro the next 20 years.... I can't imagine why we would need 128bit applications,, the 64 bit address space is "huge".....So that shouldn't be a problem. We don't really need Yet Another File System, so that should be OK.. The majority of people don't even use more than 5% of what is already available to them anyway....
Let's concentrate on security ( bye bye NSA), go back to keeoing things simpe, compatibility between platforms, anything other than the endless race to have the highest version and the maximum incompatibility.
[Please tell me that that screenshot is not using the default color scheme. I have not yet had the opportunity to actually see 10].
"Actually we do need a new filesystem."
I'd rather they just fixed (reversed) a stupid UI change when using explorer, regardless of the underlying filing system.
If I double click on a directory with the left mouse button and no modifiers, I am trying to open that directory, and that directory alone: The first click should select the directory and deselect any others, and the second should open it.
This was fine up until Windows XP - but in Windows 7 (and possibly in Vista, but I can't be sure), the first click didn't deselect anything else. So if I happen to have selected another directory when clicking into the window to make it active, double clicking another opens both the intended one, and the one previously selected.
It's easy to get around, but it's a stupid UI change.
"This happens on my Windows 8.1 machine, and happened on my previous Windows 7 machine - and in both cases, that's out of the box."
I've never sat down to explore exactly when it happens, because it's not all the time - so I've done that now...
I have it set up so that double clicking a directory opens that directory in a new window. Sometimes, I don't want the new window (or, more precisely, I no longer need to keep the parent open), so I use a modifier - CTRL - to open the directory in the same window.
It's when I do that that the problem occurs - if another directory is already selected, the one I want is opened in the same window, the other one that's selected is opened in a new window.
So, in essence, it seems that when the CTRL key is held down, Windows Explorer isn't using the first click to select the directory I want, and deselect any others - which is what it does if CTRL isn't pressed.
Changing the setting to open new windows in the same window by default, again it works as expected without CTRL - if another directory happens to be selected, it's deselected on the first click, and the directory I want is opened in the same window.
However, if I hold the CTRL key down to force the directory to open in a new window (perhaps I want to keep the parent open as well), then if another directory happens to be selected, both that one and the one I want are opened in new windows.
So Explorer is getting it wrong if CTRL is used to modify the same window/new window behaviour; it's not deselecting other items.
In your shop, when you upgrade security by changing the code in a program that has a version number, do you not increment the version number?
Most of MS's recent versions, security is what they're about. You change the release and version numbers as you make coding security updates.
And this is especially important for major security updates because major security updates almost always introduce some degree of incompatibility with poorly written programs.
Many poorly written programs, and even some well written programs, depend on OS features that violate security.
It isn't just OS X, Windows, Linux, Apache, etc. that have this issue. The internet architecture itself is incompatible with security and to become anywhere near secure against non-state actors would require such massive architectural changes that the new protocols and new code (including most firmware and software) would be incompatible with the old.
And for those others in sales and manufacturing, the ones who want the OS to have some wild new feature to sell their products:
1. An OS isn't supposed to sell products.
2. An OS is supposed to run the products that do the selling.
Desktops, lap tops, smart phones, tablets, their sales are not in the doldrums because of Windows.
Their sales are in the doldrums because of the lack of new hardware that can do something useful that the old hardware can't.
Blame Windows 7? If there was something that for most people and most companies was better that is what people would be running. I mean, Windows is cheap for OEMs, but Linux is free.
The only reason Linux hasn't wiped out Windows is it isn't as good for most customers.
And OS X, why don't governments and banks all run it? Because it is less secure against custom targeted attacks than Windows 7.
The thing that is wrong with Windows and MS isn't so much the products (although no software is perfect), it is the marketing, the PR and the management of expectations.
In those 3 things MS lags far far far behind Apple and even Linux.
"The only reason Linux hasn't wiped out Windows is it isn't as good for most customers."
Actually, for the 90% of customers that want to browse the web, send an email and maybe write a letter... It's perfectly fine. You can do all those things with little to no understanding of the underlying OS.
The only reason Linux hasn't wiped out Windows is that customers are too scared of change (which also happens to be a major factor in Windows 8's failure).
You build 128-bit applications not because of the address space but because of the increase in bus speed throughput. With a 128 bit bus you load twice as much data to the CPU in one cycle as well as increasing the speed at which the CPU processes the data..
Now, you can argue we've already passed the point at which loading data to the CPU and/or SPU processing speed has a meaningful effect to the average punter. I hold that opinion myself. But I recognize the theoretical advantage of the higher bus throughput as well as the fact that there exists a market for which that increase is useful.
You can build computers with 128-bit data buses without increasing the width of the address bus, and people have been doing that for a long time.
Most modern computer architectures - including intel and ARM - have a set of wide registers used for floating-point and SIMD aka "Media" processing. Register widths of 128-bits and 256-bits are common.The contents of the registers can be treated as multiple 8-bit, 16-bit, 32-bit, or 64-bit data items, etc. to allow parallel operations (e.g. treat a 128-bit register as four 32-bit floating-point values, and do four parallel adds to the four values in another 128-bit register and write four results into a third 128-bit register).
So the industry is already using wider buses and wider registers to get more performance.
Note that there is no need for the register widths and the bus widths to be the same. Different price performance points can be implemented that can run the same code at different performance levels. Also in the example above, different price performance points can implement different numbers of floating-point adders. Four operating in parallel gives maximum performance. One doing each of the four adds sequentially gives lowest cost.
These machines with wide buses and wide media registers are still considered 64-bit machines if they hold addresses (pointers) in 64-bit registers and have 64-bit scalar integer math and logical operations.
Changing a label only means the label has changed, the rest is pure speculation until someone gets their hands on the kernel & checks it.
If a tin of beans changes the look of its label will that mean it will have a new recipe? It might. But it's pointless guessing until you either try the beans or read the ingredients.
I was surprised they didn't do this earlier... xxx or xxxx is the build number... Windows NT 3.1 used 3.1.xxx kernel vesion; NT 3.5 used 3.5.xxxx and 3.51 3.51.xxxx kernel versions. NT 4.0 used 4.0.xxxx kernel version, and Windows 2000 5.0.xxxx kernel version. This is quite sensible. THEN:
Server 2003, XP 64-bit 5.2.xxxx
Windows 7 6.1.xxxx
Windows 8 6.2.xxxx
Windows 8.1 6.3.xxxx
So, I guess since "XP", "2003", and "Vista" aren't really numerical version numbers anyway.. whatever. But I really don't know why.... wwhhhhhhhhyyyyy..... they didn't have WIndows 7 have a 7.0.xxxx kernel, given that the previous version already had a 6.0.xxxx kernel. Windows 8 then could have had an 8.0.xxxx kernel and 8.1 8.1.xxxx. Oh well, giving Windows 10 a 10.0.xxxx kernel, better late than never.
But XP is a finished version of 2K
Server 2003 is the Server version of XP
So logical that they are 5.0 5.1 and 5.1
Vista was new level of Stupid, so 6.0
Win7 should have been a free upgrade to Vista users, it's fixed version of Vista so 6.1 is logical, Win 7 is a lie as a name. It's a fixed version of Win 6.
Win8 though ?? The numbering suggests only a GUI level Change.
Look at version numbers of Word and Office for Windows since Word 2.0 ! Massive jumps.
Why has Apple OS 10 stayed at 10 since introduction? Are they really only incremental changes in last 15 years? I don't think so. I think they liked the Roman Numeral X
Was there really 9 versions of Mac OS before they scrapped it and ported Next's flavour of BSD?
Mage is absolutely correct - Vista (NT 6.0) was something NEW (Complete with a new set of faults/bugs). "Windows 7" (NT 6.1 was a major (and very successful) bug fix. "Windows 8" was indeed just a GUI change because if you install a new shell such as Classic Shell, it can work EXACTLY like what it is - NT 6.1 (Windows 7).
They charged for "Windows 7" to extract more money out of the muggins who paid for NT 6.0 (We who have been in the industry for some time know: NEVER, EVER buy x.0 from Microsoft - it never works properly).
It's probably by now means particularly slow or bad in any way. It does have some interesting ideas like "Personas" which would allow you to have different APIs.
What people complain about is the Windows user space. It stops supporting old software and drops vital features while gaining irrelevant ones. The user space is where most security bugs lie. If you'd just install cygwin directly on top of Windows, you would probably have a rather decent and secure operating system.
What most people complain about is inability to support existing peripherals after an upgrade. If a new kernel requires new drivers, then what are the odds that the manufacturer of your 5=year-old flat-bed scanner will write an updated one?
re. flat bed scanner: I was annoyed when I found that Windows 7 wouldn't talk to my Epson Perfection 1650 scanner. I had to keep my old XP laptop going to use it. However, the drivers are available for Linux and that was only one of the many reasons I swtiched to Linux about 18 months ago.
You don't have to go far. Services For Unix currently found in Windows Programs & Features/Turn Windows Features On. Something I've had installed since forever. Still works for in 8.1 although they've deprecated it. Again.
It's where I go to csh.
FTA: Microsoft's effort to unite the desktop, server, and mobile versions of its OS around a single, common core
I'll say this for Microsoft: when they've got something in their sights, they keep shooting until it dies, no matter how painful it is, and no matter if the thing in their sights turns out to be their own foot.
At least they finally blinked and gave up on their "New Coke" moment of forcing the stupid Metro tiles down everyone's throat. 8.1 is a big improvement over the GUI-as-a-religion 8.0 nonsense; now 8.1 is just like the Mormons constantly knocking on your door offering tiles.
And I understand they finally fired the idiot that was pushing it so hard.
Maybe MS should do what Apple does and keep its OS's secret from the press until they're ready for prime time.
The press just never seems to tire of its amazement that a pre-beta not even aurora release of something is not a finished product.
MS has already shown their Windows versioning system to be upper marketing bollocks.
8, 8.1, 10, then probably freaking 20 and so on.
It doesn't come as a surprise the kernel versioning system is going the same way ! Always the same code, with a crapped on UI, but V20.0 vs. V2.1, people will pay for it, yay ! Same goes with IE, from 6.0 to 11.0 ... And patches going for all 6 versions ...
Frankly, move along, nothing to see here ...
Can I get rid of that crap at the bottom of the screen, unpin all that shite from the taskbar, and does it shove me into happy-clappy-bouncy-animated-icons-your-PC-is-now-a-phone land when I press the start button?
Also can I use it and keep it up to date without a Microsoft account?
If I find the answers favourable, I might get it after various people have made a WGA un-fucker for it. If not, oh well, I'll stay with 7 until end of life, and then the Linux partition becomes the permanent primary.
Technology / business experts in several countries have speculated that the 10 designation of "new" NT based Windows has a great deal to do with satisfying the Chinese government and markets - about 500 million users - that this is a non-NSA subverted Operating System - in contrast to banned Windows 8, that will live up to a credible level of reliability and security.
It will be interesting to see if the Windows 10 for Desktops and Tablets will share the substantially improved "ResFS" File System used in Servers, but totally improbable - technically and practically - for implementation in Windows phones, in publicized campaign of “One Windows”.
However the 'tested' inferiority of ResFS in comparison to ZFS and btrfs File Systems used in *NIX Operating Systems won't change the dynamic of Windows Server not being competitive or a "viable" challenge in overall Cloud Computing, Virtualization with Containerization, and Social Media/Enterprise infrastructure and Networking implementations worldwide.
For those of us old enough, we can remember them wanting that back in NT3.5 days.
When they brought out Windows 2000, there was only one desktop version and only one server version.
It was brilliant. I had the same version on my home PC, work and a voluntary organisation I supported
Licensing was not quite so over-complicated. All software worked on all PCs (hardware permitting) and they hadn't started to prat around with the UI.
Presumably, someone worked out that they had less room to gouge customers and the wildly proliferating versions started to come back with XP.
It would be nice to go back to just one desktop version of Windows but I suspect that they have other plans - and not because it will benefit users either.
Biting the hand that feeds IT © 1998–2020