andrew@jung.harlqn.uucp (Andrew Watson) (10/15/88)
In article <485@genesis.ATT.COM> andys@genesis.ATT.COM (a.b.sherman) writes: In article <ANDREW.88Oct5134036@jung.harlqn.uucp> I write: >Can anyone give me a good reason why all these big corporations licence System >V, only to have to port to it large parts of 4.1/2/3 just to make it usable? >I'm talking about sockets, the Fast File System, all the TCP/IP utilities, >symbolic links ... and then they're *still* left with a system with a lousy >short term scheduler that truncates all file names to 14 characters! > I don't really want to get into a debate about the technical merits of these comments. However, people license System V and port it because it is the latest commercially available version of the UNIX system. Also, they will have to pay the license fee for most other *IX systems anyway, since they were mostly derived from AT&T source code of older versions. Okay, I'll get off my hobby horse now, before I start another skirmish in the Unix religious wars (:-) ... I've used both System V.1/2/3 and BSD 4.2 (vanilla, SunOS, Ultrix) extensively, and honestly prefer the latter, and I'm still curious to know the reasons behind the decisions of the likes of HP, DG, TI, Intergraph &c to go with System V not 4.2. Is it really just the licencing minefield surrounding the Berkeley version? Can someone cast some light on this? -- Regards, Andrew. +-----------------------------------------------------------------------------+ | Andrew Watson, Harlequin Limited, andrew@uk.co.harlqn | | Barrington Hall, Barrington, Tel: +44 223 872522 | | Cambridge CB2 5RG, UK Fax: +44 223 872519 | +-----------------------------------------------------------------------------+
peter@thirdi.UUCP (Peter Rowell) (10/18/88)
In a previous article, andrew@jung.harlqn.uucp (Andrew Watson) writes: >I've used both System V.1/2/3 and BSD 4.2 (vanilla, SunOS, Ultrix) extensively, >and honestly prefer the latter, and I'm still curious to know the reasons >behind the decisions of the likes of HP, DG, TI, Intergraph &c to go with >System V not 4.2. Is it really just the licencing minefield surrounding the >Berkeley version? Can someone cast some light on this? Actually it's quite simple - once you accept that not all decisions are based on purely technical considerations. In this case, the various manufacturers must decide to offer either BSD (maintained by "whoever hasn't graduated yet") or System V (maintained by "The Inventors of UNIX and a major Fortune 50 corporation"). The existence of BSD and SysV has created a type of tension. SysV has given apparent stability to the code and real credibility to the future of UNIX in the "real world". BSD has been what university projects have always been - where the cowboys hang out. They can get more than a little rowdy, try out a new dance step or two, and if they break the mirror in the back of the bar..... welllllll boys (and girls?) will be boys..... The funny thing is, I believe the BSD releases have been more solid at initial release than the System V ("commercial quality") stuff has been. The upcoming System V.4 may very well mark the end of this basic tension. AT&T is finally admitting the "kids" from Berkely might have a few good ideas and the people at Berkeley South (== Sun) are accepting that they need the increased commercial credibility created by kissing GrandMa in public. Never fear - with OSF, POSIX, X/Open, etc. we will have plenty of "friendly" debate to keep things interesting and (hopefully) moving forwards (or sideways or ???). "The great thing about UNIX standards is that there are so many to choose from!" ---------------------------------------------------------------------- Peter Rowell Third Eye Software, Inc. (415) 321-0967 Menlo Park, CA 94025 ...!pyramid!thirdi!peter
guy@auspex.UUCP (Guy Harris) (10/19/88)
>I've used both System V.1/2/3 and BSD 4.2 (vanilla, SunOS, Ultrix) extensively, >and honestly prefer the latter, and I'm still curious to know the reasons >behind the decisions of the likes of HP, DG, TI, Intergraph &c to go with >System V not 4.2. Define "go with System V". For example, as I understand it, many (most? all?) HP-UX implementations are based on at least a BSD kernel, although the fast file system used to truncate names to 14 characters anyway (I've heard you can now ask it not to) and a lot of the user-mode code and much additional kernel code probably comes from System V. By contrast, SunOS implementations are based on at least a BSD kernel and a lot of the user-mode code and much additional code comes from System V. I think the same applies to Ultrix. (In all three cases, there's a lot of HP, Sun, and DEC code in the OS as well. SunOS 4.0's "mmap" sure didn't come from Berkeley....) Lots of UNIXes these days are built from some System V code, some BSD code, and some code from other sources. I'd much rather work on a system with stuff from both than a pure version of either one. >Is it really just the licencing minefield surrounding the >Berkeley version? To what "licensing minefield" do you refer? I suspect you can't get a 32V license any more, so you have to get an S5 license to get BSD source, but so what?