8

AT&T released UNIX Version 7 (seven) in 1979.

The same company released UNIX System V (five) in 1983.

Why did the later release have a lower number?

DrSheldon
  • 15,979
  • 5
  • 49
  • 113
  • 7
    Of course, Microsoft is even worse at counting: 1, 2, 3, 95, 4, 98, 2000, 7, 8, 10. – DrSheldon Dec 19 '20 at 01:53
  • 3
    Those are Windows marketing names. winver shows the version number just fine. – RonJohn Dec 19 '20 at 05:47
  • 1
    They are basically two different OSes released by different companies (so no, not the same company) – slebetman Dec 19 '20 at 14:51
  • 2
    @RonJohn Actually, winver on my system shows the version as 2004 (plus a build number). Typing [environment]::OSVersion.Version in PowerShell shows the major version as 10, but Windows 8 would I believe show 6. So whichever way you slice it, they've jumped a few. – IMSoP Dec 19 '20 at 17:54
  • @IMSoP Looks like they changed something along the way, kinda like Chrome and then Firefox changed their naming scheme. (Winver on Server 2008 Ent says that the version is "6.0 (Build 6003)". Windows Server 2016 says "Version 1607 (OS Build 14393.22)". Still, newer versions have highe numbers. – RonJohn Dec 19 '20 at 19:40
  • 2
    Microsoft skipped Windows 9 intentionally, because too many apps used a check for "Windows 9" and refused to run because they thought it was Windows 95 or Windows 98. – Mark Ransom Dec 23 '20 at 03:19
  • @MarkRansom "too many apps" means Java, but what I never understood was: those version strings came from within Java itself so surely they could have fixed the check at the same time they added the new version string for Windows 10. – user253751 Apr 05 '22 at 11:21
  • @user253751 one is still too many. The check wasn't in Windows itself, so there was nothing they could do to fix it. – Mark Ransom Apr 05 '22 at 12:50
  • @MarkRansom but until people updated Java, it would have said something like "Windows unknown" and after updating Java, "Windows 9" and the bug would also be fixed – user253751 Apr 05 '22 at 12:53
  • @user253751 remember that Microsoft was sued for anti-competitive practices. Breaking Java and waiting for them to fix it would have been seen as evil by far too many. – Mark Ransom Apr 05 '22 at 13:07
  • @MarkRansom except you were either using pre-update Java, in which case you weren't broken, or post-update Java, in which case you weren't broken – user253751 Apr 05 '22 at 15:32

1 Answers1

19

There is not a single Unix line ('Unix' is not unique). The numbers measure different things. Anyone who forked a variant of Unix was free to start a whole new sequence of numbering.

Seventh edition Unix came from the 'classic' Unix lineage out of Bell Labs, latterly known as 'Research Unix'.

Meanwhile, presumably because of its flexibility, there were many offshoots from the original, under various divisions or departments. In particular for this answer, the Unix Support Group (USG) developed a commercialized Unix system. System III contained features from several different Unixes: 7th ed. Unix, PWB/UNIX 2.0, CB UNIX 3.0, UNIX/RT and UNIX/32V.

Per Wikipedia, the system was apparently called System III because it was considered the outside release of UNIX/TS 3.0.1 and CB UNIX 3.

System V followed on from System III (System IV never seems to have emerged in public).

But to state my answer differently, there's no reason why the number sequence of the commercial "System N" variants should follow on from the numbers used by Bell Lab's "N'th edition" Unix.

dave
  • 35,301
  • 3
  • 80
  • 160