>>>>>>I've been a Microsoft fan for decades, but I'm working hard to jump ship at this point.
>>>>>
>>>>>What are you jumping to?
>>>>
>>>>That's the tough part. No really good alternatives.
>>>
>>>The toughest and in the end unbeaten competitor to the Vista was XP. W7 will be the same for W8.
>>
>>
>>Is there pattern to it ?
>>They usually come with one which is all right right after the one that is considered rubbish. Maybe they will come again with something better with W9 ?
>
>Remember DOS 4.0? I do, not fondly at all. That POS even managed to have duplicate directory entries. W95 was returning 4.0 on Os(), but it was officially an odd number... and most of its troubles were fixed in W98, aka 4.1... the Vista is 6.0, and then W7 is actually 6.1.
>
>The pattern is "any M$ OS with even number and dot zero is crap".

Some disasters I remember with DOS 4.0 (Microsoft blames IBM for these)
* Dos-shell -- more like dos-hell, rather kludgy w/ .COM, .EXE and .BAT components
* Default install assumes new install -- it formatted drive w/o prompting
* Incompatible FAT format (for partition > 32MB) -- especially deadly if you used certain disk tools (e.g. Norton Defrag)

And let's not forget the DOS 6.0 disc compression -- it's OK as long as DON'T use Windows 2.x or 3.x with it.

One thing that I *did not like* about DOS 6.0 was the backup and restore utilities -- it was changed so you were forced to eject and insert floppies while the drive was still running. With 5.25" minifloppy it was generally OK -- the eject lever would physically disengage the read-write head before ejecting. With 3.5" microfloppy drives it was a different matter -- the drive head wasn't physically disengaged before the disc was ejected -- so while the disc was being spun, the head was resting on the disc. Since the eject button didn't physically disengage the read-write head, you basically drove the disc up against the read-write head -- eventually you ended up misaligning or damaging the read-write head.

Of course, when it came to getting version of OS, it was tricky in Win16 -- there was the Windows version, and there was the underlying DOS version. Though officially Win95 was Windows 4.0, it reports 3.95 -- apparently due to some programs having broken version check (I suspect there may have been some Microsoft products amongst them). That sounds like a silly explanation -- but if you take a look at the low-level API, you find a problem -- like DOS, the major version is returned in low-order byte and minor release in high-order byte -- meaning you had to be careful with version check. The following type of code would fail unexpectedly:

if ( WindowsVersion() < 0x3003 ) then  // make sure we're using at least v 3.30
    messagebox "You need to upgrade Windows"
    halt
endif

One thing that always bugged me about OS version was when you had utilities like DOSVER - which would configure the OS to lie. Though understandably this was added to get certain programs to work properly, it meant it was sometimes a pain to get the *real* version.