On Tue, 2006-05-23 at 15:07 +0100, Kevanf1 wrote:
:-) is there really a wrong reason for using Linux?
Yes! Unless you embrace the concept of open standards, Linux/Open Source is _not_ an ideal solution. Vendors will _never_ give companies a "way out" of proprietary lock-in. Yes, Linux/Open Source will often let you move away from _older_ proprietary software -- but attempting to "stay current" on proprietary software while attempting to adopt Linux/Open Source has resulted in some _utter_failures_ I've seen. Another thing that Linux is oversold on is eliminating Enterprise Configuration Management (ECM) costs. Yes, UNIX/Linux _reduces_ ECM costs over Windows, but it does not eliminate them. The problem is that Linux is oversold on entry costs, not realizing there are still significant ECM overheads just like Windows. This was the same problem at NASA with Quality Assurance (QA) slashing in the '90s. Mars Pathfinder cut hardware and software development costs by 90%. So the attitude was to cut the entire budget by 90%, which included QA. So it did not surprise me one damn bit when the Mars Polar Lander came in shallow because two different teams across two different contractors were using two different metrics in their code. I've really seen some consultants oversell Linux and Open Source on things that are unavoidable. And it really makes Linux and Open Source look bad -- when it was poor planning.
Hmm.... I'll await the howls of protest but surely not. We are talking about the PC here not a workstation linked to a mainframe.
No, there are _assumptions_ that come from the Windows world! Multi-user concepts are the big-ass #1. Package and dependencies are the big-ass #2. And so forth! Hell, there are even _key_ application differences. E.g., Windows executives _require_ a "start-up directory" and that's a major pain. Then there are the attitudes and automation that the Windows executive has that UNIX/Linux will _never_ adopt because they violate basic multi-user security (let alone in the age of the Internet). You really need my class man! ;->
The mainframe/workstation set up is different to the original PC that came from IBM. I know that there were earlier incarnations of the small form factor computer but for all intent and purpose the PC as we know it came about because of IBM's offerings. IBM first used a DOS for those PC's.
DOS was not only single user, but single task! DPMI and Windows added task-swapping, but was still single user. In fact, Citrix hacked multi-user into the NT kernel, but it's still not the same as UNIX. _Huge_ "shifts" in thinking!
Now, feel free to correct me but didn't Microsoft either write that software or come in very soon after and started to write the software?
Seattle Computer Products ported CP/M from the 8080 to 8086 without a license. Microsoft licensed that for $50,000. IBM settled the license issue out of court less than a year later for $800,000. MS-DOS 2.0 legally obtained SCO Xenix code to augment file, directory and streamed I/O operations. IBM and legally obtained Apple code created Windows, as well as the early DOS Protected Mode Interface (DPMI) services, ultimately resulting in Protected286/386 modes. OS/2 and Windows split the design -- with OS/2 adopting a full Protected386 mode, while Windows used (including in MS-DOS 7.x/Windows 4.x "Chicago" 95/98/Me) 386Enhanced which constantly shunts the processor between Real86 and Protected386. Microsoft continued to sap IBM OS/2 code even after the expiration of the 1981 agreement in 1993 -- and IBM's PC division signed away its IP rights in 1995 just before Windows 95 hit (long story). Windows NT, based on the OS/2 API and codebase, was not only written by former Digital VMS developers, but Digital continued to help _heavily_ in its development (long story). This included the current Win64 API. Anyone who ran Digital software on Windows NT could instantly tell the level of quality different in actual use of the NT/Win32 API versus those "designed for Chicago" from Microsoft.
So, really, the Microsoft way is actually the initial way of having an operating system.
You are surely joking, correct? The concept of _true_ multi-user -- originally the "time share" operating system -- was invented _before_ UNIX. UNIX just made it commodity.
Linux, as a Unix way of doing things came about a few years later.
GNU began in the early '80s as a fundamental licensing change due to AT&T's monopoly being busted and it now being allowed to assert IP rights over existing UNIX developments. The "GNU System" begins _before_ Linux and _heavily_ influences its design! MIT's "W" windowing environment begins _before_ Windows' earliest developments, even along Apple's! The concept of "remote display" begins with MIT-Digital with "X" in the early '80s. Let alone Xerox was almost a decade before that! KEY POINT: EVEN THE ORIGINAL MOUSE IN 1965 HAD *3* BUTTONS! Xerox's prototype has 3 buttons. MIT-Digital W and X have 3 buttons. Almost _everything_ we used today was invented by Xerox and Digital, with key points thanx to AT&T.
Please note that nowhere have I said the Microsoft way is the correct way :-) Just because they were first does not mean they were, or indeed, are correct in their method of running an operating system on the PC. I believe the *nix way is far better. I imagine IBM couldn't afford the then available Unix on their new range of PC's because of prohibitive costs?
SCO Xenix existing in 1979, and gave you an UNIX platform for 8086 -- *2* years before the IBM PC. Microsoft had a 10% investment in SCO Xenix from day 1, including rights to the code -- which they used _heavily_ when they augmented MS-DOS 2.0 from virtually its original "CP/M pirated" PC-DOS 1.0. Remember, the 8086 came out in 1976 -- 5 _years_ before the PC. There were other UNIX flavors for Intel 8080 and 8086 (including the latter/cheaper 8088) from the late '70s through early '80s. It was only the proliferation of the IBM PC, and their failure to enforce patents on companies like Compaq, that led to mass adoption. IBM originally approached Digital Research to do CP/M for 8086/8088. When Digital Research showed no interest, they approached Microsoft who was doing their ROM Basic to write one. Luckily they knew about Seattle Computer Products illegal rip of CP/M and rebuild for 8086. Understand Gates was a _major_pirate_ himself! He _knew_ other piraters! But back then, they were called "developers" and "hackers." Bill Gates wrote his infamous 1975 letter "most of you steal your software" from a 100% _hypocritical_ standpoint. He bitched about fellow students taking his code and using it, _failing_ to mention that his version was _ripped_ from others and just slightly modified. Even Microsoft Basic sold to Altair and others was based on Digital's (Digital Equipment Corporation, DEC) Basic. CASE-IN-POINT: *NONE* OF MICROSOFT'S PRODUCTS WERE "TIME SHARE" AND "TRUE MULTIUSER" -- LET ALONE WITH PRIVILEGE LEVEL (before NT -- although NT suffered from "Chicago" and the lack of privileged level). A big "buzzword" today in the Windows world is "least privilege model." We call that "standard" in UNIX. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com ----------------------------------------------------------- Americans don't get upset because citizens in some foreign nations can burn the American flag -- Americans get upset because citizens in those same nations can't burn their own