Today Microsoft launched its latest version of its iconic Operating System: Windows 8. There have been many version of Windows over the years, and one can’t help but wonder what justifies this version number. Microsoft and other technology companies have over-written and re-named their product history before, so we take a good hard look at all the previous operating systems to see if the numbers add up.
Microsoft Windows started as version 1.0 as an add-on to the DOS operating system. The product was a simple GUI program that was little more than a visual file manager. Windows 2.0 was released to take advantage of the features offered by Intel’s 286 processor. Windows 3 is the version that really started Microsoft on its current path of success.
All of these versions numbers are very logical and can be easily counted. Where things start to get a little ambiguous is when Windows NT is thrown into the mix.
On the next page we start at the simple and humble beginning…
Version numbers are augmented all the time, which are usually set by marketing departments because the the average person assumes that a higher version number is just better. It happens all the time. The second-generation Xbox was competing with the Playstation 3, but Microsoft couldn’t just name their new console the “Xbox 3” with a straight face, so they named it the “360” because it had a “3” in the name. Subliminally, the public thinks they are now comparing the two competing consoles on a level playing field.
The same thing happened with Microsoft Word. Microsoft Word was trying to take down the most popular word processor on the PC at the time, which was called WordPerfect, who had just released their 6th version. Word was re-branded Microsoft Word 6.0, even though it was actually their third version of the word processor at the time. Again, this was done because the average person in the software store believes that “version 6” of something is obviously better than “version 3” of a competing product.
What happened to Word versions 3.0, 4.0, and 5.0?
This type of thing continued on with its Windows products, and the confusion became much worse when Windows NT was introduced to the public.
Windows NT confounded the simple version system.
Windows NT (short for New Technology) debuted its version number as 3.1 (the same as Windows 3.1). Windows NT was superior to Windows 3.1 in nearly every way, and was marketed to professionals and corporations who wanted a real, reliable operating system. Unfortunately, marketing Windows NT 1.0 would have been a sales disaster, as the public would think that Windows 3.1 was the better product. Windows NT 3.1 was short-lived, and the new Windows NT 3.5 was marketed to differentiate it as a “better” version of Windows 3.1, even though it was a completely different operating system
There are certain logical rules when defining version numbers. A product that is written to support a new architecture gets a new version number. A product that has a lot of new features gets a new version number, such as Windows 3.0. A minor enhancement gets a subversion, as did Windows 3.1.
If we keep to these same logical rules, then we can count up what the “real” version number of Windows truly is.
There was a time when Windows wanted to get away from the whole “version number” thing to avoid confusion, and went with years. That way, people would understand that Windows 98 was a newer version than Windows 95, although they both had pretty much the same code base.
The thing to remember is that DOS-based Windows (Windows 95, 98, Me, etc) and NT-based Windows (NT 4.0, Windows 2000, etc.) were marketed at the same time as completely different operating systems for completely different demographics. Windows NT 4.0 had the same look and feel of Windows 95, but was based on a completely different architecture. The “consumer” version of Windows now was associated with a year, and the “professional” version had a version number. Pretty simple to understand.
Different versions for different hardware?
Another thing to throw into the mix is that there were different versions of Windows NT 4.0 for RISC processors. Usually when you write an OS for a different hardware architecture, it’s a different version, right? That’s what Microsoft did when they wrote Windows 3.0; they created a new operating system to support the new 386 processor. Should Windows NT 4.0 for RISC be a subversion of NT4 or should it have its own version number? We’ll consider this when we add everything up later.
Then Windows 2000 came along. This was the technological achievement that Microsoft had been trying to do for years: kill off unstable DOS code with an OS that was compatible with all existing software. Windows 2000 played DirectX games, which was something that only Windows 9x would previously do.
Windows Me (Millennium Edition) was released shortly after Windows 2000, and was the last version to be based on 16-bit DOS code. The average person was confused, because they didn’t understand that one was a professional version, and another the consumer version.
The thing to remember is that Windows 2000 is actually NT version 5.0, and Windows Me is a subversion of Windows 95. Since Windows 95 was the successor to Windows 3.1, then shouldn’t W95 be Windows 4.0? Wait, there was already a Windows NT 4.0. Do you see where this starts getting a little confusing?
Windows 2000 was also built for 64-bit Itanium processor. Going by the rules of version numbers, should this operating system have an incremented version number, or be considered a subversion (or flavor) of Windows 2000 (NT 5)?
The successor to Windows 2000 was Windows XP (the Windows Experience), and was even more accessible to the average person than Windows 2000. If Windows 2000 is NT 5.0, then XP has to be 6.0, right? Forget about that whole Windows 95/98/98SE/Me should be version 4.x even when Windows NT 4.0 already existed.
During the Windows XP time, Microsoft also released different flavors for Media Center PCs (XP Media Center Edition). This is all well and good, but they released a new version of it every year (MCE 2003, MCE 2004, and MCE 2005).
The next version of Windows was Windows Vista, which Microsoft would like to pretend never existed. According to our existing NT tally, Windows Vista should be NT version 7. Wait, isn’t Windows 7 version 7? Is Microsoft trying to pawn off that Windows Vista was just a “preview” of Windows 7? Or do they consider Vista to be a subversion of XP (NT version 6)?
The Server Confusion
If all of this isn’t confusing enough, we have to consider the Server flavors of all of these Operating Systems. Instead of naming the Server edition of XP “Windows XP Server” (like they did with Windows NT 4.0), Microsoft dubbed it “Windows Server 2003”. This distanced the name from XP (the consumer version, which was really the professional codebase) and their Server line (the professional version). Should the Server line be considered a subversion of XP, even though Microsoft obviously marketed it as a completely different version?
64-bit versions
Last but not least, we also have to consider the 64-bit versions of the operating systems. By the Rules of Versioning, any new hardware architecture gets its own version number. 64-bit versions of Windows XP supported the (relatively) new AMD 64 processors, and since then all general-purpose PC processors are 64-bit capable. But at the time the 64-bit version of XP was created to support a whole new hardware architecture. Should it get its own version number? What about the 64-bit versions of Vista, Windows 7, and Windows 8?
What about Windows 8 RT? The RT version only supports the ARM processor inside Microsoft’s new Surface tablet, which again is a completely different hardware architecture.
In order to count what the “true” version number of Windows 8 is, we’ve laid out a handy chart with all of the major releases of Windows. As we’ve pointed out earlier in this article, it is debatable whether some versions should be considered its own version number or be lumped in as a subversion of an existing edition.
Microsoft was not consistent with how it marketed some of its versions. Windows 98 was marketed as a new version of Windows, even though it was really a subversion of Windows 95. The same thing is true with Server 2003, even though it was a subversion of XP. In the opposite direction, Microsoft named Windows NT 3.5, trying to make the public think it was in the same product line as Windows 3.1.
Will the real 3.1 please stand up?
So, whenever a break in the rules occurs, we mark the version as a potential increase and count it as a “Version Bump”. Whenever a version of Windows supports a new hardware architecture, it also gets a “Version Bump”.
We can now calculate the true version of Windows 8.
| version | OS Name | subversions | version bump |
| 1 | Windows 1.0 | 1.01, 1.02, 1.03, 1.04 | – |
| 2 | Windows 2.0 | 2.03, 2.10, 2.11 | – |
| 3 | Windows 3.0 | 3.1, Windows for Workgroups 3.11* | +1 |
| 4 | Windows NT 3.1 | Windows NT 3.51* | +1 |
| 5 | Windows 95 | OSR1, OSR2, USB Supplement, OSR 2.1, OSR 2.5, Windows 98*, 98 SE, Windows Me* |
+2 |
| 6 | Windows NT 4.0 | 4.0 Server*, Terminal Server, NT 4.0 for RISC** |
+1 |
| 7 | Windows 2000 | 2000 Server*, Itanium Edition** | +2 |
| 8 | Windows XP | Starter, Home, Home N, Professional, Professional N, Tablet PC, MCE 2003*, MCE 2004*, MCE 2005*, Tablet Edition, Server 2003*, x64** |
+5 |
| 9 | Windows Vista | Starter, Home Basic, Home Premium, Business, Enterprise, Ultimate, x64 versions*, Server 2005* |
+2 |
| 10 | Windows 7 | Starter, Home Basic, Home Premium, Professional, Enterprise, Ultimate, Server 2008*, Server 2008 R2, x64** |
+2 |
| 11 | Windows 8 | Pro, Enterprise, x64**, RT** | +2 |
* = subversion significant enough or marketed as new OS
** = different hardware architecture
The chart looks pretty straightforward? Not so fast… read on…
When everything is laid out, Windows 8 is obviously not version 8. If Microsoft wants to re-write its history of version numbers, then it is admitting post-mortem that some versions were irrelevant and that customers did not need to buy the “new” OS.
The only way Microsoft can say that Windows 8 is truly version 8 is if they break the rules and re-write the history of their version numbers. To whittle down to this version number you have to really broaden your definition of what a version number is and accept the following illogical conditions as true:
- Lump all subversions and different hardware architectures in the same version number
- Merge Windows 95 in with NT 4.0 (making all Windows 9x as 4.x)
- Merge Vista with Windows 7 (admitting that Vista was a “pre-release”)
- Make the leap that Windows NT 3.1 and 3.51 were a subversion of Windows 3.0
If you actually count all the version bumps: Windows for Workgroups 3.11, Windows NT 3.51, all the versions MCE and Server as seperate Operating Systems, then the number goes way up.
If you count x64 versions as a different OS (using the same rules we applied to not mix 16-bit and 32-bit OSes), Windows 8 RT (ARM-based), Windows 2000 x64 (Itanium) then the number jumps even higher.
So what is the real version of Windows 8?
That depends on how you look at it.
Conservatively, Windows 8 is actually Windows 11.
If you count all the subversions that should be a new version in its own right (according to how Microsoft marketed them), then Windows 8 is really version 24.
If you count all the subversions and the different versions for hardware, then you are on Windows 29.
Regardless of how Microsoft wants to re-label and market their software, there were certain best-of-breed versions that stand out as major milestones. These versions were Windows 3.1, Windows NT 4.0, Windows 98 SE, Windows XP, and Windows 7. The jury is still out on how Windows 8 will be received.