That's an explicit reference to PF Zachary's book and PF makes no such allegation. Brett Alexander Hunter ( talk) 07:41, 23 October 2021 (UTC)
I'm new to Wikipedia so I couldn't work out how to change this. But in the infobox on the left-hand side, under the "Latest stable release", there is some boldface text proclaiming that Windows is "bad", and that Ubuntu is far better. Whilst I may or may not agree with the sentiment, this looks like vandalism to me. Qlexander ( talk) 10:31, 23 April 2013 (UTC)
I don't know the origins of NT well enough to write even a good stub article. I simply neutralised the comments that existed before, and took out the BSoD reference entirely -- Colin dellow
The first sentence is not even English. This article, from the point of view of the informed user needing more information, is largely useless. Needs to be edited to bring out the facts (release dates, versions, features, which parts of the technology have been carried into Win2K and XP) and remove most of the opinions, which help no one other than the writer. --BK
There are a number of mistakes regarding the history of MS-Windows in this and related articles. First up NT, was NOT based on OS/2 v.3 which was released a few years after NT. Nor was OS/2 writen by Microsoft, it was a joint IBM/Microsoft project to produce a mature OS for the home PC market, just do a search for a famous video clip of Bill Gates in 1989 cooing over how OS/2 is going to be the OS of the 1990's. It was only after he saw the sales figures for a semi workable GUI environment that he realised selling vapour ware would be more profitable than a joint mature & working product which he'd have to share with IBM, who's engineering skills well deserved. To this end Mr Gates had his people do a universal text-processor change of the source code for OS/2 ver. 1.2 (it may have been 1.3, but I think it was 1.2) replacing the tag 'DOS' with 'WIN'; recompiled, and called that Windows-NT version 1.0 . This was the reason that NT at first 'could' run OS/2 programs, because they couldn't stop it from doing so for the first half dozen releases. You must remember OS/2 1.x was a 16 bit OS, IBM re-wrote from the ground up as a pure 32 bit OS for ver. 2.0 which had the Windows 3.0 'emulator'; which was easy as 90% of the Windows code was written by IBM code cutters under contract & that contract which gave IBM access rights was the reason that IBM also wrote bug-fixes for MS-Windows 3.0 & 3.1; however, Microsoft took IBM to court to prevent IBM from releasing the fixes & claiming IBM had no rights to any of the Windows code. The court case was bogus but MS kept in in the courts from 1992-1995 when it of course became meaningless.
Basicaly, Windows NT is the Microsoft version of OS/2.
MS-Windows95 on the other hand was a Desktop system mimicing the OS/2 Desktop system - complete with weaknesses as well as strengths- sitting on top of MS-Dos v.7 The modern Microsoft Windows is a Microsoft version of the OS/2 Desktop manager on top of a Microsoft version of OS/2. And OS/2 was a fine OS for its time, and for its market; but it nor anything based from it is NOT suitable for todays network environment.
IBM didn't "produce" OS/2; Microsoft did. It was released as Microsoft OS/2 1.0, a complete operating system, entirely produced by Microsoft. IBM bought it, improved upon it, and released it under their name later. My impression at the time was that MS used it as something of a "testbed" for how to write an OS, then pawned it off to IBM and started over with NT. Internally, most of the MS team who produced OS/2 were the same folks who produced NT (For example, the OS/2 component test team directly became the NT component test team, which was where I spent my brief employment at MS). --
LDC
As I recall it, it was a joint venture between the two of them. I'm sure each side would claim it did the lion's share. I recall that significant amounts of IBM code went into OS/2, including in particular the graphics system. There was plenty of disagreement, because IBM wanted it to fit into their SAA architecture, while Microsoft wanted to keep compatibility with the Win16 GDI. In any case, I know specifically that OS/2 versions 1.3 and 2.0 were released under both Microsoft and IBM brands. That simply could not have happened if IBM just bought it later. The schism happened after OS/2 2.0; IBM went on to release OS/2 3.0 and Microsoft released NT 3.0. That's why there was no NT 1.0 or 2.0
-- Alan Millar
I'm sure you're right about IBM being more involved; I used Microsoft OS/2 1.0, but I didn't join MS until after the team had moved to NT, so I can't speak to the details in between. I can, however, state with some confidence that the reason the first release of NT was called "3.0" was to synchronize its version number with that of the currently-popular Windows 3.0. By the way, if you have an old copy of NT 3.1 lying around, you'll find the email addresses of the team, including "leecr", in the easter egg. --LDC
Agreed. Ive actually got a copy of Nt3.1. It was superceded by 3.5 and 3.51. It uses program manager like windows 3.1 but clearly says Windows NT on it. Its got networking, and is a fully 32 bit os. It can still run some modern programs, in fact. 134.36.92.18 ( talk) 00:06, 26 April 2009 (UTC)
Although I don't have it direct from Cutler, I hardly knew him and he knew me even less, I am as certain as I can be that the VMS => WNT ploy was intentional. Cutler was fond of puns and word play and worked a number of jokes into RSX, where I worked with him. See Talk:RSX-11 for some examples that are verified out of my own experience and some indications as to his personality. He was a profane and sarcastic individual who felt unappreciated by DEC and was glad to go to Microsoft. It would be absolutely like him to rub it in. So no, the VMS=> WNT thing isn't verified, but I know it happened. Ortolan88 20:47 Aug 4, 2002 (PDT)
This article was added to
Category:Discontinued Microsoft software this morning. However, is that really an accurate classification for NT?
On the one hand, Microsoft has not marketed new products in this line of software under the trademark "NT" since the release of Windows 2000. On the other hand, each release of Windows since that time has reported itself as, inter alia, "Windows NT 5.0" (in Windows 2000), "Windows NT 5.1" (in Windows XP) or the like, which suggests that Microsoft continues to use the appelation "NT" at least internally.
—
Ryanaxp 14:02, Jun 14, 2005 (UTC)
In lots of internal things, microsoft products still use NT. E.g lots of active directory binaries use the acronym NTDS (NT Directory Service) 134.36.92.18 ( talk) 00:07, 26 April 2009 (UTC)
It should be undiscontinued. MS frequently refer to the current windows kernel architechture as the NT architechture in developer docs —Preceding unsigned comment added by 134.36.93.46 ( talk) 03:03, 21 May 2009 (UTC)
NT also has several different flavours not mentioned here, I think. I know that Terminal (Server) Edition also exists for NT 4.0. Advanced Server and Server exist in Windows 2000. Windows 2003 has at least Enterprise Edition. Jdstroy 03:07, 2004 Aug 13 (UTC)
Why no mention of NT 4.5? Edit: for those who claim there's no such thing, Windows NT 4.5 Small Business Server and Windows NT 4.5 BackOffice Server.
What part am I not explaining. They take a pre-existing version of Windows Server (originally Windows NT 4.0 Server, then Windows 2000 Server and now Windows Server 2003), then add on a bunch of programs (Exchange Server, Proxy Server, etc), stick them all in a box and sell them. It is not a separate version of Windows. Small Business Server is a separate product, not a version of Windows. All it has just now is a redirect Microsoft Small Business Server, please feel free to expand on that if you want. AlistairMcMillan 00:37, 2 September 2005 (UTC)
Windows 2000 has a datacenter variant. There is NT embedded, windows 2000 advanced server, nt4 terminal server edition (joint program with citrix i think). And BackOffice server "NT 4.5" is just NT4 with exchange and mssql etc... bundled. 134.36.92.18 ( talk) 00:09, 26 April 2009 (UTC)
Does anyone have a source for this...
This was added by User:216.119.235.214 over a month ago. [1] I'd love to know if this is true. AlistairMcMillan 22:23, 30 August 2005 (UTC)
Someone recently deleted the mention of Windows NT as an example operating system on the Mach Kernel page. I've seen no evidence that NT uses Mach, and at best only have seen comparisons and hints at inspiration from Mach. I'm deleting the mention of Mach from the "See also" section. -- 64.222.109.153 22:17, 19 March 2006 (UTC)
Windows XP Professional x64 Edition is branded as Windows XP (NT 5.1), but based on Windows Server 2003 (NT 5.2). Which NT version number does Windows XP Professional x64 Edition claim to be? -- Pmsyyz 23:33, 17 September 2005 (UTC)
Yeah its 5.2. Also the most recent service pack for it is SP2 (uses its owne modified server 2003 serivoce packs)
134.36.92.18 (
talk)
00:10, 26 April 2009 (UTC)
http://search.microsoft.com/search/results.aspx?st=b&na=88&View=en-us&qu=longhorn Microsoft sites say Vista has replaced longhorn. 71.28.243.246 19:58, 13 October 2005 (UTC)
Vista has not been released yet (as of
21 February
2006) and is expected around the end of 2006 (although it could slip into 2007 - my opinion). The article has now been reverted the to reflect this again. Microsoft has info about the Vista beta
[2]
Imagine_B
10:43, 21 February 2006 (UTC)
I reverted an anonymous edit that changed disk to disc. Looking at the wiktionary definition of disk, disc is the North American variant of disk; using this description it is best to keep the terminology consistent. There are roughly three occurances of disk so changing one of them to disc is not a good idea.
The other way to look at it, is in international English (according to wiktionary definition of disc) disc implies "optical disk", in which case hard drive space does not qualify to use disc. BigNate37 19:58, 26 June 2006 (UTC)
Is it necessary to make the comparisons between UNIX and Windows NT? I realize that there is a great schism between the masses of Windows NT and UNIX users, but I don't feel that it would be appropriate to make the claims of:
I'd like to point out that, up to Windows NT 5.0, there *is* a POSIX subsystem (posix.exe and pax.exe) deployed on a standard installation; thus, Windows NT 5.0 and earlier had built-in support for a UNIX platform. This doesn't really fit the description of "features comparable to UNIX," as it contains a UNIX platform. Windows NT's "subsystem" architecture was designed in a manner similar to the idea of "universes" on a UNIX system--it allowed a user to run programs in multiple platforms, including the Win32 subsystem (and 16-bit DOS/Win16 subsystem), the POSIX subsystem, and the OS2 subsystem.
In later versions of Windows NT (5.1, 5.2), the support for the POSIX subsystem was dropped; however, an add-on package from Interix (and now Microsoft) called "Windows Services for Unix" allows for the use of the POSIX subsystem on Windows NT 5.1 and Windows NT 5.2.
Although Windows NT 5.0 and earlier had support for the POSIX subsystem in a standard installation, it lacked the tools that went along with a standard UNIX install. (It's like having a Linux kernel, but no GNU!) The rest of the runtime environment was left in the Windows Services for Unix package.
I believe the second quote (above) is incorrect. SFU is "a POSIX.1-conforming system" [3]; thus, it has source compatibility (up to the POSIX.1 specification, see POSIX) with "UNIX."
Thoughts? Jdstroy 04:36, 4 July 2006 (UTC)
Its purpose was to have a complete implementation of Posix that was absolutely no use to code against, which was achieved by not offering any network API, no graphics API, or anything other than a command line. Why do it? So people purchasing NT boxes on government deals could check off the "Posix compliant" box, which was often a requirement.
SteveLoughran 21:24, 9 January 2007 (UTC)
It may be ineresting to note that MS bundle a number of unix utilities with windows. In the windows release notes its got the disclaimer (the one BSD type licences require you to have). I read that utilities like FTP were rebuilt by MS from sources from the OpenBSD project. —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:47, 25 April 2009 (UTC)
Since increases in processor performance, Windows XP and Windows Server 2003 are actually the latest versions of Windows NT, though they are not branded as such for marketing purposes.
How does the fact that Windows XP and Windows Server 2003 are not branded as NT follow from the fact that CPUs are now faster?
"Released versions of NT for Alpha were 32-bit only"
This doesn't really make sense and I don't think it is even possible. The Alpha is purely a 64-bit architecture. -- Afed 16:55, 18 August 2006 (UTC)
Indeed. The shipped versions of NT for Alpha were pure 32-bit operating systems and supported only 32-bit applications. Same for the first versions of VMS for Alpha, for that matter.
It's true that the chip has no 32-bit mode. This is handled through what amounts to sign extension: Any memory reference on an Alpha has to be done by loading the pointer into a register. And any code compiled for 32-bit Alpha, when loading a pointer into a register, treats it as a signed number, copying bit 31 of the pointer into bits 32-63! Thus the 32-bit OSs on the Alpha populate the very first 2 GiB and the very last 2 Gib of the 64-bit address space, in a manner very similar to the 48-bit "canonical" virtual addresses on X64.
However, unlike 32-bit mode on e.g. X64, all 64 bits of the Alpha's GPRs (and all of the GPRs, for that matter) were available for "large integer" calculations in this "32 bit" environment!
MS released 32-bit NT for Alpha through NT version 4, and shipped some 32-bit Alpha beta versions of Win2K. This continued after CPQ dropped support on NT for Alpha, but with no intention on MS's part of shipping the code. It was done for two reasons: 1) There was Alpha hardware but no Itanium hardware; and 2) they wanted to maintain a code base for more than one processor architecture, a practice that made the (much later) X64 port much easier. But no 64-bit NT code was ever shipped, even in beta form. (I assure you that if it had been, I'd be running a copy!) I imagine that now that they have three architectures to support the Alpha branches have not seen further development. (the above shamelessly plagiarized from my own text on the talk page of Architecture of Windows NT ) Jeh ( talk) 03:23, 14 January 2008 (UTC)
The article says:
What does "to complement workstation versions of Windows that were based on MS-DOS until 2001" mean here? I'm not sure I'd call the DOS-based versions of Windows "workstation versions" - NT was better suited to " workstations" than the DOS-based versions. (The workstation article speaks of UNIX, but that's arguably an error; machines used as scientific and engineering workstations often run NT.)
I'd be inclined to say that NT was originally mainly a workstation and server version of Windows, complementing the DOS-based lower-end consumer and business desktop versions of Windows. It's now replaced the DOS-based versions in the latter market. Guy Harris 00:05, 19 August 2006 (UTC)
Aha, so 'NT' is expanded to 'New Technology' for marketing purposes. And also for marketing purposes Windows 2000 had the slogan 'Built on NT technology.' So this expands to 'Built on New Technology technology.' This must be a joke, right? iNic 02:40, 23 November 2006 (UTC)
I do expect marketing to make logical sense. If the marketing doesn't make logical sense why would their products make logical sense? For example, marketing for New Age products doesn't make logical sense—but neither do their products. So that is fine with me. But maybe MS products doesn't make logical sense either? In that case I will rest my case. (BTW, I've never heard anything but PIN code, never PIN number.) iNic 02:25, 25 November 2006 (UTC)
NT is not about New Technology, but about a HAL9000 like joke... If you take Windows NT, remove the "indows" you will have WNT, and if you do a ROT-1 you will get "VMS" (it's much like "HAL" in "2001: A Space Odyssey", if you to a ROT1 you will get "IBM") - Godzil -- 212.157.49.189 14:45, 26 September 2007 (UTC)
The NT repetitive thing is outlined in this article for any of you out there who want to read about it. Emprovision ( talk) 19:12, 27 May 2008 (UTC)
Like Unix, NT was written in C, a high level language. It can be recompiled to run on other processor systems, at the expense of larger and slower code.
Two problems with this statement: first, what's the basis of the comparison? "Other", "larger", "slower" compared to... Intel x86? Intel i860? MIPS R4000?
More important: A priori there's no reason to believe the claim that recompiling on another plaform gives larger and slower code compared to the original. Without a reference and clarification, this statement should be removed. Bhudson 21:02, 6 January 2007 (UTC)
-If i remember rightly, when NT war originally made (nt 3.x) the dev machines used were deliberately not intel x86 machines. The reason behind this was to stop the developers using nonportable hacks. —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:51, 25 April 2009 (UTC)
Alistair: WinME was a bastard operating system that was released in mid-1999, because MS could not get NT5 out the door in time for the Christmas 1999 season. NT5, rechristened Windows 2000 at the last minute by the marketroids, was released in a big coast-to-coast bash featuring Carlos Santana in mid-February 2000.
By the way, I was a beta tester for both the x86 and AlphaNT versions of Win2k... In fact, I still run RC2 (build 2128) on my DEC AlphaPC 164SX. Dan Schwartz Discpad 23:22, 13 March 2007 (UTC)
Yeah MS wanted to combine the home (DOS) and business (NT) lines into one (like they did eventually with xp) in tje late 90s, but Windows Neptune (codename of the planned OS) was not ready for home use just yet, so they released win2k and ME as a stopgap (me was made in like 6 months).
Thats why XP came out only like 9 months after windows 2000, and they are very similar. Windows XP was basically windows 2000 + the home stuff like the application compatibility wizard, network setup wizard etc.... planned for neptune.
If you try the windows neptune alpha (some on the net) you will see what i mean. Its basically a very buggy win2k, with some wizards that didnt emerge until xp (like the beginnings of a welcome screen implementation) —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:54, 25 April 2009 (UTC)
The article says Home Server will be NT 5.2. If there is a source to this claim please post here if not I will change it to:
NT Ver. | Marketing Name | Editions | Release Date | RTM Build |
---|---|---|---|---|
Unknown | Windows Home Server | Unknown | 2007 (expected) | Unknown |
Chetblong 20:03, 1 June 2007 (UTC)
I went ahead and changed it to Unknown. -- :) Chetblong 19:02, 7 June 2007 (UTC)
It is 5.2. I own it. Its a version of 2003 Small Business Server with domain functionality and IIS removed, among other things + addons like the new volume manager and backup.
The file version it comes up as (on nlite etc...) is 5.2
[all nt <vista can be installed to fat..] "at the expense of speed and security" no, fat is faster than NTFS for normal use on all but the very largest drives.~blab
;Only on small partitions running DOS! Per KB: "over 200 MB the FAT file system should not be used. This is because as the size of the volume increases, performance with FAT will quickly decrease." per http://support.microsoft.com/kb/100108 . Also consider problems like more than a few thousand entries in a directory (like Internet Cache).
"Limitations of FAT32 File System". Microsoft Help and Support. 2004-12-16. http://support.microsoft.com/kb/184006/en-us Chen, Raymond (2006). Microsoft TechNet: A Brief and Incomplete History of FAT32. TechNet Magazine July 2006 http://groups.google.com/group/microsoft.public.win98.fat32/browse_thread/thread/1fa5783ca7a26a5b
Shjacks45 ( talk) 13:20, 13 April 2010 (UTC)
The Hardware requirements section doesn't mention year difrences between the respective operating systems, something I feel may make it be unbiased against Microsoft for the percieved software bloat (noted, they rely less on application optimization and more on Moore's Law in order to get speed increases). If the years were listed in the table, perhaps it would be more effective at showing the data? - 68.228.56.158 19:13, 20 July 2007 (UTC)
Which is the most stable NT version? I know the kernel is refined all along the way, and NT 5 (Windows 2000) included a major architecture redesign. However, what about the whole system's buglessness and stability? In Yahoo! answers one said it was Windows 2000 that was the most stable version. I myself felt that too (patched at least up to year 2004). In my experience XP is also stable, however with some strange user mode hang-up situations: when switching users, logging off, etc. -- 211.167.159.70 08:06, 14 August 2007 (UTC)
-the server builds are the most stable IMHO. Server 2000 or 2003 have never crashed for me under general use, unlike XP and 2000 Workstation —Preceding unsigned comment added by 86.16.160.17 ( talk) 21:45, 2 January 2008 (UTC)
Depends on what you consider stable. Downtime? In which case probably Server 2008 - its got self-healing NTFS and hotpatching. Windows 2000 is mature - but that doesnt nececerially mean stable. —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:58, 25 April 2009 (UTC)
Does the controversy concerning the minute different versions of NT Server (4.0) and NT Workstation (4.0) deserve to be brought up in this article? http://www.oreilly.com/news/differences_nt.html —Preceding unsigned comment added by 75.74.240.92 ( talk) 02:56, 27 November 2007 (UTC)
Release list should include Home Server and processor type for all releases. Hel pslo ose 21:08, 20 December 2007 (UTC)
NT was written in C, a mid-level language. This means that it can be compiled to run on several processor systems; however, the code produced by the compiler is larger and slower than assembler code written for a particular processor[citation needed]. For this reason, NT was not favored initially for use with slower processors with less memory.
As a CS engineering student, I can almost certainly say this is wrong. Code produced by a compiler is not inherently slower, and it will CERTAINLY not be the reason why a certain OS cannot be used for a certain machine. Heck, Linux avoids ASM code where slightly possible, only very lowlevel function that cannot be described in C will be programmed in ASM. —Preceding unsigned comment added by 193.190.253.149 ( talk) 15:52, 11 January 2008 (UTC)
What it means is that NT is slower than 9x.
9x had lots and lots of asm code, which was highly optimesed for x86. —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:59, 25 April 2009 (UTC)
What do you think of the table now? I confess I'd forgotten completely about XP 64-bit Ver 2003! (and MS did too, very soon after...) I have to run out now, but later I'll put in a link to that press release.
yeah, there is XP x64 (Intel Itanium architechtute) and XP Pro x64 (the common x86_64)
XP x64 (itanium) was pure xp SP1. XP Pro x64 was based on Server 2003 x64
Do you think the supported processor architectures should perhaps have a column of their own? This might mean a few more new rows, but it's arguably useful to have it in the table. Jeh ( talk) 03:15, 14 January 2008 (UTC)
XP 64 bit (Intel Itanium) was based on pure xp, not 2003. It was only supported up to SP1 i think. XP ~Pro~ x64 is supported up to SP2 and is based on Server 2003. Dont get them confused.
134.36.92.18 ( talk) 00:12, 26 April 2009 (UTC)
Do we even have a source for this? If not, it should be removed. I doubt Windows 7 will be NT 6.1, especially considering it's Windows 7, which comes after Windows NT 6/Vista. 99.149.127.80 ( talk) 06:43, 5 April 2008 (UTC)
MS said (in a webcast on msdn discussing windows driver development for windows 7 for OEMs) that windows 7 will be NT 6.1, as will windows server 2008 R2. This is in order to keep compatibility apparently. —Preceding unsigned comment added by 134.36.92.18 ( talk) 00:01, 26 April 2009 (UTC)
I doubt it also that it will be NT 6.1, it's not logical, and further more my I heard that Microsoft named 'Windows 7' Windows 7 because they want to go back to the naming of their OS with the version number... Like they did in the beginning... Windows NT 3.1, Windows NT 3.5, Windows NT 3.51, etc... Wouser ( talk) 13:08, 18 May 2009 (UTC)
It is DEFINATELY 6.1 thats what it says in the windows 7 release candidate and what i read in some prelease material aimed at OEM driver developers. So will be Server 2008 R2 (which will be based on windows 7, as opposed to the original server 2008 which was vista sp1). Windows '7' is purely a marketing term. it is NOT nt 7. —Preceding unsigned comment added by 134.36.93.46 ( talk) 03:07, 21 May 2009 (UTC)
There's precedent for the official marketed version number and the actual OS-reported version number being different: Windows for Workgroups 3.11 actually reports its version as "3.10". --20:47, 4 May 2011 (UTC) —Preceding unsigned comment added by 69.116.117.48 ( talk)
Windows Server developer Preview uses 6.2 Kernel Winver. — Preceding unsigned comment added by 79.237.254.50 ( talk) 10:56, 8 February 2012 (UTC)
I think this source is certainly more reliable than this one. Accordingly, if no one has any objections, the Xbox line should be removed. Or at least reworded so that it says the Xbox OS is derived from the Win32 APIs but built from the ground up. - xpclient Talk 11:41, 22 May 2008 (UTC)
Windows 7 is still a codename and not a "marketing name", Windows 7 is schedueld to be released in H1 2010, not H2 2009. —Preceding unsigned comment added by 129.241.150.119 ( talk) 15:55, 30 August 2008 (UTC)
Windows 7 WILL be sold as "Windows 7". Its the release/marketing name now, not just a codename. —Preceding unsigned comment added by 134.36.93.46 ( talk) 07:13, 22 May 2009 (UTC)
See this. I don't feel like creating an account to fix this/cite that part of the article. Check the second paragraph. —Preceding unsigned comment added by 69.84.252.227 ( talk) 00:28, 31 January 2009 (UTC)
There's some disagreement regarding whether the TCP/IP implementation is derived or imitating BSD. However, Winsock is derived from BSD. Tedickey ( talk) 14:06, 4 April 2010 (UTC)
see xbox, dreamcast I worked at Microsoft in 1997 when NT 5.0 beta code was forked to Xbox. No Win32 or other subsystems. Windows Native Mode, Drivers, and DirectX; all kernel mode stuff. Not sure of Xbox 360, but there was PowerPC NT 5.0 code available at the time (except for Internet Explorer). Dreamcast was WinCE with DirectX, bare bones CE will fit in "250K" but not directx. Most modern BIOS is 128KB (to 512KB+) which includes chipset setup and EFI boot code then shrinks to 64KB after boot. Fact is most games load their operating system from the game CD or game cartridge, see Playstation "disk swap" section. There was a project there to "secure boot" PCs by putting MBR code and boot files into ROM (Flash) to prevent "viruses". NTLoader, NTDetect, boot.ini et al can fit into 250K. PicoXP is only 14MB. Shjacks45 ( talk) 14:48, 13 April 2010 (UTC)
"Initial development was for the Intel i860 RISC then MIPS until i386 became available.[5][6] ". According to the wikipedia articles about these processors the i860 was produced from 1989 and the i386 from 1985. The references 5 and 6 are mostly circular and are not useful. —Preceding unsigned comment added by L00KnS33 ( talk • contribs) 07:48, 11 August 2010 (UTC)
Windows 2000 64-bit was codenamed "Janus".
So how come it's not on Windows NT#64-bit? Also see Talk:Windows 2000#64-bit. - 79.179.200.231 ( talk) 19:34, 9 January 2011 (UTC)
Here's an old Microsoft page about it:
WhisperToMe ( talk) 01:58, 14 October 2011 (UTC)
Steve Kleynhans said: "Windows 8 is more than a major upgrade to Windows - it's a technology shift. We don't see technology shifts too often; the only other one Microsoft's client OS has gone through was the move from DOS technology to Windows NT technology, which began in 1993 and took eight years, ending with the introduction of Windows XP in 2001,"
News link: http://www.ibtimes.com/articles/356436/20120626/windows-8-marks-beginning-winrt-winnt-gartner.htm — Preceding unsigned comment added by 77.121.210.102 ( talk) 15:38, 3 July 2012 (UTC)
This also uses an NT 6.x kernel. Should it be mentioned? PizzaMan ( talk) 12:59, 26 May 2013 (UTC)
The Page is Currently Locked and Contains Information that is Uncited and I Know is incorrect. Windows 3.1 Didn't Run on the NT Kernel it ran on the DOS kernel and I know this for a fact. Someone Should Correct this I think it's more than a simple mistake.
The quote as given is not suitable for the technical description, since the point of it was to accuse competitors of being dishonest. There's no technical issue presented in the comment. TEDickey ( talk) 13:16, 13 October 2013 (UTC)
I would like to see a citation supporting the following claim:
"Windows NT is a family of operating systems produced by Microsoft, the first version of which was released in July 1993 ... Windows 2000 and its successors are members of the Windows NT family, although they are not branded using the name 'Windows NT'." and for the inclusion of Windows 8.1.
I don't see any justification for calling every version of Windows until the end of time "Windows NT". I would have less of a problem if we differentiated between Windows NT the operating system and the Windows NT family of operating systems and if we used Category:Windows NT family of operating systems or even Category:Windows NT architecture instead of Category:Windows NT, but most of the references I can find for "Windows NT family of operating systems" are rather old and in the context of explaining how this new OS is different from the Windows 95 family of operating systems. I would very much like a citation to a reliable source that tells us which operating systems are or are not part of the Windows NT family of operating systems. -- Guy Macon ( talk) 22:16, 22 April 2014 (UTC)
Guy Macon, Jeh, Dogmaticeclectic, Guy Harris, Codename Lisa, I decided to tackle the issue of finding a WP:RS that ties all of this into a "family". Microsoft's OSINFO structure's article says:
The use of the word "descended" is still vague on if we can call this a "family" but is getting closer. Microsoft's OS_TYPE Enumeration links the family a bit better in that it classifies machines as one of:
Click on <224> above to learn that a "Windows NT workstation" indicates "Windows 2000 Professional, Windows XP, Windows Vista, Windows 7, Windows 8, or Windows 8.1."
Click on <226> above to learn that a "Windows NT server" indicates "Windows 2000 Server, Windows Server 2003, Windows Server 2008, Windows Server 2008 R2, Windows Server 2012, or Windows Server 2012 R2."
I had also looked at the WIN32 API's GetVersionEx() function which returns an OSVERSIONINFO structure that contains the dwPlatformId field. This field is defined as holding the value 2 or VER_PLATFORM_WIN32_NT meaning "The operating system is Windows 7, Windows Server 2008, Windows Vista, Windows Server 2003, Windows XP, or Windows 2000." While that Microsoft article talks about Windows 8 and 8.1 it's silent on what the dwPlatformId is for those operating systems. I've submitted feedback to Microsoft to see if they will fix that detail.
I suspect Microsoft's OS_TYPE article and its footnotes are decent WP:RS that Windows NT and its descendants up through Windows 8.1 / Windows Server 2012 R2 are in the same family. I'm not sure how to work that into the Wikipedia article without a bunch of hand waving. I'd like to "show" rather than "tell" the WP reader it's a family.
In my own mind I've always called it the "Windows NT architecture" rather than a "family" but in reading over this thread it seems "family" is a better word and that we use "Windows NT" to separate this from the older generations of the Microsoft Windows family. -- Marc Kupper| talk 22:20, 11 September 2014 (UTC)
The result of the move request was: not moved. -- BrownHairedGirl (talk) • ( contribs) 17:54, 8 May 2014 (UTC)
Windows NT →
Windows NT family (typo corrected; was "Windows NT Family") – This rename will differentiate between Windows NT the specific operating systems named "Windows NT" and Windows NT the family of related operating systems that includes Windows XP, Windows 7, and Windows 8. --
Guy Macon (
talk)
20:55, 26 April 2014 (UTC) Edited 12:04, 27 April 2014 (UTC)
*'''Support'''
or *'''Oppose'''
, then sign your comment with ~~~~
. Since
polling is not a substitute for discussion, please explain your reasons, taking into account
Wikipedia's policy on article titles.Recently all versions of Windows (XP+) have been changed to Windows N.T. ¿Should Windows Phone 8 and later also have this on the top of their infoboxes? as they are also based on the Windows N.T. kernel ¿or are they different because of other reasons? Sincerely, -- Namlong618 ( talk) 16:45, 18 February 2015 (UTC)
The Windows NT family includes mobile, embedded, desktop/notebook, and server OSes. The infobox's release information is for the desktop/notebook flavor; should there be a separate infobox for the server flavor, and perhaps also on the Microsoft Windows page, or should that go on the Windows Server page? This page doesn't give a marketing target in the infobox; the Microsoft Windows page does, giving it as "personal computing", but it also covers non-desktop/notebook flavors of Windows. Guy Harris ( talk) 23:20, 19 July 2015 (UTC)
Current version says that the 10-fold disk space increase in Vista is primarily due to the extra memory needed by the Desktop Window Manager. This doesn't sound right -- it's mixing system memory / VRAM with disk space. The Vista DWM was notorious for using a lot of memory and causing slow performance as a result -- which was fixed by rearchitecting it in Windows 7 -- but I've never heard of it being a significant cause of disk space usage and no citation has been provided for this claim. 24.130.133.67 ( talk) 09:18, 22 July 2015 (UTC)
Windows 10 is actually 6.4 despite the 2014 news stating that it will be 10.0. Someone should clean up the article. ebertek ( talk) 15:53, 24 March 2016 (UTC)
Shouldn't there be a mention that NT was bought from Northern Telecom? — Preceding unsigned comment added by 67.78.5.254 ( talk) 15:45, 8 April 2016 (UTC)
There is currently a discussion of whether Template:Infobox OS should be used with multiple version numbers - for example, to list both a "software update" and "next major release" beta, or to list betas from more than one release stream. If you believe that multiple {stable, preview} releases should never appear in that infobox, or if you believe that they should appear under some or all circumstances where there's more than one beta of the OS in question available, you might want to comment there. (I have no strong belief either way; I'm OK with the main OS page listing only the "next major release" beta, but listing betas from multiple streams if they exist, but I'd also be OK with other choices.) Guy Harris ( talk) 08:09, 11 July 2016 (UTC)
Please see Talk:Windows NT 4.0 § NT 4 kernel source code on GitHub — Codename Lisa ( talk) 07:07, 21 May 2017 (UTC)
Not that I want move this page to a new title, but is there any reason it is still move protected over a single instance of page move vandalism from 8 years ago? It doesn't seem to meet the purposes set out in WP:MOVP. Kb.au ( talk) 14:42, 30 June 2017 (UTC)
This article is such a page. In applying all Wikipedia policies, common sense is a requirement. In this case, the decision to protect it is just good sense.Highly visible pages that have no reason to be moved
I think the following paragraph (last in the overview section) uses encyclopedic language, and could be unnecessarily promotional, but I didn't want to make changes since I don't often visit this page and am not familiar with the history:
"In some ways the ancestors of Windows NT are the operating systems that Dave Cutler worked on before he was hired by Microsoft, namely the VMS and RSX-11 operating systems, and also an unreleased object-based operating system developed by him for DEC Prism. This lineage is made clear in Cutler's foreword to "Inside Windows NT" by Helen Custer.[4]"
The phrase "in some ways" is clearly nonspecific and those "ways" should probably be enumerated. Also, I think the general tone of this sounds a little promotional towards Dave Cutler (he is the author of the source cited to prove the claim that he developed the "ancestors" of Windows NT. If he is really a big figure in Windows NT that's obviously fine, but this sentence doesn't make clear what his contributions were, or why they were important enough that he is listed in the overview section rather than the history section.
Basically, when reading this section I felt kind of like I was talking to some weathered Computer Scientist giving me his version of how Windows came to be, not like I was reading an objective encyclopedia. — Preceding unsigned comment added by Ludicous ( talk • contribs) 16:45, 8 September 2017 (UTC)
Why the new Windows logo is being used for this 1993 released OS? BarricadeX ( talk) 04:00, 14 August 2018 (UTC)
This page should not be speedy deleted as pure vandalism or a blatant hoax, because I WANNA DELETE THIS PAGE — Preceding unsigned comment added by Ziadelwan ( talk • contribs) 19:08, 29 August 2018 (UTC)
The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for deletion:
Participate in the deletion discussion at the nomination page. — Community Tech bot ( talk) 10:40, 30 June 2021 (UTC)
2401:D800:2220:8C36:3152:2512:9718:BA30 ( talk · contribs · WHOIS) / 2401:d800:2de0:e28e:56fd:f712:2d14:f7fa ( talk · contribs · WHOIS) (I'm assuming they're the same person) has recently tagged this page – and hybrid kernel – with {{ Citation needed}}, stating that we need official confirmation from Microsoft to call NT a hybrid kernel. I don't see how that is grounded in policy, though. WP:V requires reliable sources, so if reliable sources call NT a hybrid kernel, that should be stated. I don't know of any policy that would additionally require confirmation by the vendor of a product in a case like this. ;; Maddy ♥︎(they/she)♥︎ :: talk 08:30, 24 August 2022 (UTC)
The idea behind a hybrid kernel is to have a kernel structure similar to that of a microkernel, but to implement that structure in the manner of a monolithic kernel. In contrast to a microkernel, all (or nearly all) operating system services in a hybrid kernel are still in kernel space. There are none of the reliability benefits of having services in user space, as with a microkernel. However, just as with an ordinary monolithic kernel, there is none of the performance overhead for message passing and context switching between kernel and user mode that normally comes with a microkernel.
}The Windows NT 4.0 architecture merges the best attributes of a layered operating system with those of a client/server or microkernel operating system. [2] The Windows NT Executive is the generic name for a number of subsystems and components of the operating system that execute in the microprocessor's kernel (or privileged*)* mode. The remainder of the operating system, such as various server subsystems and their application clients, execute in user (or application) mode.
A microkernel, on the other hand, is the name given to the core portion of a modern, modular operating system. Microkernel operating systems are based on two fundamental principles. The most basic principle is one of modularity, encapsulation, and data hiding. In this aspect of the design, there is one and only one portion of the operating system that has system-wide responsibility for a particular function. All other parts of the operating system (as well as applications, naturally) access that function through a well-defined interface. There is no duplication of function and no "back doors" to critical data structures; all access to system data structures is through software interfaces. This approach makes it possible to upgrade or replace entire modules within the system without disturbing the remainder. A secondary principal of the microkernel design, related to the first but focused more on the implementation strategy, is that large portions of the operating system which traditionally run entirely in the kernel or privileged mode of the microprocessor can now be executed in user or application mode, with only the microkernel itself, along with a relatively small amount of hardware device-specific code, executing in kernel mode.
CreateFile()
sending a message to CSRSS, which then constructs a path name that it then passes to the NtCreateFile()
system call, the result of which it sends back.)open()
is a system call, unlike CreateFile()
, and fork()
, the exec
calls, and posix_spawn()
are system calls, unlike CreateProcess()
. I'm not sure how its graphics/window system layer currently looks, but I suspect it may not be too different from current Linux systems using
Wayland. With DriverKit, some drivers - including, from my experience with Ethernet driver crashes when running
macOS Ventura on a
VMware Fusion virtual machine, at least one Ethernet driver - may have been moved from the kernel to user mode processes.References
As to the whole "hybrid kernel" thing - it's just marketing. It's "Oh, those microkernels had good PR, how can we try to get good PR for our working kernel? Oh, I know, let's use a cool name and try to imply that it has all the PR advantages that that other system has.
That's an explicit reference to PF Zachary's book and PF makes no such allegation. Brett Alexander Hunter ( talk) 07:41, 23 October 2021 (UTC)
I'm new to Wikipedia so I couldn't work out how to change this. But in the infobox on the left-hand side, under the "Latest stable release", there is some boldface text proclaiming that Windows is "bad", and that Ubuntu is far better. Whilst I may or may not agree with the sentiment, this looks like vandalism to me. Qlexander ( talk) 10:31, 23 April 2013 (UTC)
I don't know the origins of NT well enough to write even a good stub article. I simply neutralised the comments that existed before, and took out the BSoD reference entirely -- Colin dellow
The first sentence is not even English. This article, from the point of view of the informed user needing more information, is largely useless. Needs to be edited to bring out the facts (release dates, versions, features, which parts of the technology have been carried into Win2K and XP) and remove most of the opinions, which help no one other than the writer. --BK
There are a number of mistakes regarding the history of MS-Windows in this and related articles. First up NT, was NOT based on OS/2 v.3 which was released a few years after NT. Nor was OS/2 writen by Microsoft, it was a joint IBM/Microsoft project to produce a mature OS for the home PC market, just do a search for a famous video clip of Bill Gates in 1989 cooing over how OS/2 is going to be the OS of the 1990's. It was only after he saw the sales figures for a semi workable GUI environment that he realised selling vapour ware would be more profitable than a joint mature & working product which he'd have to share with IBM, who's engineering skills well deserved. To this end Mr Gates had his people do a universal text-processor change of the source code for OS/2 ver. 1.2 (it may have been 1.3, but I think it was 1.2) replacing the tag 'DOS' with 'WIN'; recompiled, and called that Windows-NT version 1.0 . This was the reason that NT at first 'could' run OS/2 programs, because they couldn't stop it from doing so for the first half dozen releases. You must remember OS/2 1.x was a 16 bit OS, IBM re-wrote from the ground up as a pure 32 bit OS for ver. 2.0 which had the Windows 3.0 'emulator'; which was easy as 90% of the Windows code was written by IBM code cutters under contract & that contract which gave IBM access rights was the reason that IBM also wrote bug-fixes for MS-Windows 3.0 & 3.1; however, Microsoft took IBM to court to prevent IBM from releasing the fixes & claiming IBM had no rights to any of the Windows code. The court case was bogus but MS kept in in the courts from 1992-1995 when it of course became meaningless.
Basicaly, Windows NT is the Microsoft version of OS/2.
MS-Windows95 on the other hand was a Desktop system mimicing the OS/2 Desktop system - complete with weaknesses as well as strengths- sitting on top of MS-Dos v.7 The modern Microsoft Windows is a Microsoft version of the OS/2 Desktop manager on top of a Microsoft version of OS/2. And OS/2 was a fine OS for its time, and for its market; but it nor anything based from it is NOT suitable for todays network environment.
IBM didn't "produce" OS/2; Microsoft did. It was released as Microsoft OS/2 1.0, a complete operating system, entirely produced by Microsoft. IBM bought it, improved upon it, and released it under their name later. My impression at the time was that MS used it as something of a "testbed" for how to write an OS, then pawned it off to IBM and started over with NT. Internally, most of the MS team who produced OS/2 were the same folks who produced NT (For example, the OS/2 component test team directly became the NT component test team, which was where I spent my brief employment at MS). --
LDC
As I recall it, it was a joint venture between the two of them. I'm sure each side would claim it did the lion's share. I recall that significant amounts of IBM code went into OS/2, including in particular the graphics system. There was plenty of disagreement, because IBM wanted it to fit into their SAA architecture, while Microsoft wanted to keep compatibility with the Win16 GDI. In any case, I know specifically that OS/2 versions 1.3 and 2.0 were released under both Microsoft and IBM brands. That simply could not have happened if IBM just bought it later. The schism happened after OS/2 2.0; IBM went on to release OS/2 3.0 and Microsoft released NT 3.0. That's why there was no NT 1.0 or 2.0
-- Alan Millar
I'm sure you're right about IBM being more involved; I used Microsoft OS/2 1.0, but I didn't join MS until after the team had moved to NT, so I can't speak to the details in between. I can, however, state with some confidence that the reason the first release of NT was called "3.0" was to synchronize its version number with that of the currently-popular Windows 3.0. By the way, if you have an old copy of NT 3.1 lying around, you'll find the email addresses of the team, including "leecr", in the easter egg. --LDC
Agreed. Ive actually got a copy of Nt3.1. It was superceded by 3.5 and 3.51. It uses program manager like windows 3.1 but clearly says Windows NT on it. Its got networking, and is a fully 32 bit os. It can still run some modern programs, in fact. 134.36.92.18 ( talk) 00:06, 26 April 2009 (UTC)
Although I don't have it direct from Cutler, I hardly knew him and he knew me even less, I am as certain as I can be that the VMS => WNT ploy was intentional. Cutler was fond of puns and word play and worked a number of jokes into RSX, where I worked with him. See Talk:RSX-11 for some examples that are verified out of my own experience and some indications as to his personality. He was a profane and sarcastic individual who felt unappreciated by DEC and was glad to go to Microsoft. It would be absolutely like him to rub it in. So no, the VMS=> WNT thing isn't verified, but I know it happened. Ortolan88 20:47 Aug 4, 2002 (PDT)
This article was added to
Category:Discontinued Microsoft software this morning. However, is that really an accurate classification for NT?
On the one hand, Microsoft has not marketed new products in this line of software under the trademark "NT" since the release of Windows 2000. On the other hand, each release of Windows since that time has reported itself as, inter alia, "Windows NT 5.0" (in Windows 2000), "Windows NT 5.1" (in Windows XP) or the like, which suggests that Microsoft continues to use the appelation "NT" at least internally.
—
Ryanaxp 14:02, Jun 14, 2005 (UTC)
In lots of internal things, microsoft products still use NT. E.g lots of active directory binaries use the acronym NTDS (NT Directory Service) 134.36.92.18 ( talk) 00:07, 26 April 2009 (UTC)
It should be undiscontinued. MS frequently refer to the current windows kernel architechture as the NT architechture in developer docs —Preceding unsigned comment added by 134.36.93.46 ( talk) 03:03, 21 May 2009 (UTC)
NT also has several different flavours not mentioned here, I think. I know that Terminal (Server) Edition also exists for NT 4.0. Advanced Server and Server exist in Windows 2000. Windows 2003 has at least Enterprise Edition. Jdstroy 03:07, 2004 Aug 13 (UTC)
Why no mention of NT 4.5? Edit: for those who claim there's no such thing, Windows NT 4.5 Small Business Server and Windows NT 4.5 BackOffice Server.
What part am I not explaining. They take a pre-existing version of Windows Server (originally Windows NT 4.0 Server, then Windows 2000 Server and now Windows Server 2003), then add on a bunch of programs (Exchange Server, Proxy Server, etc), stick them all in a box and sell them. It is not a separate version of Windows. Small Business Server is a separate product, not a version of Windows. All it has just now is a redirect Microsoft Small Business Server, please feel free to expand on that if you want. AlistairMcMillan 00:37, 2 September 2005 (UTC)
Windows 2000 has a datacenter variant. There is NT embedded, windows 2000 advanced server, nt4 terminal server edition (joint program with citrix i think). And BackOffice server "NT 4.5" is just NT4 with exchange and mssql etc... bundled. 134.36.92.18 ( talk) 00:09, 26 April 2009 (UTC)
Does anyone have a source for this...
This was added by User:216.119.235.214 over a month ago. [1] I'd love to know if this is true. AlistairMcMillan 22:23, 30 August 2005 (UTC)
Someone recently deleted the mention of Windows NT as an example operating system on the Mach Kernel page. I've seen no evidence that NT uses Mach, and at best only have seen comparisons and hints at inspiration from Mach. I'm deleting the mention of Mach from the "See also" section. -- 64.222.109.153 22:17, 19 March 2006 (UTC)
Windows XP Professional x64 Edition is branded as Windows XP (NT 5.1), but based on Windows Server 2003 (NT 5.2). Which NT version number does Windows XP Professional x64 Edition claim to be? -- Pmsyyz 23:33, 17 September 2005 (UTC)
Yeah its 5.2. Also the most recent service pack for it is SP2 (uses its owne modified server 2003 serivoce packs)
134.36.92.18 (
talk)
00:10, 26 April 2009 (UTC)
http://search.microsoft.com/search/results.aspx?st=b&na=88&View=en-us&qu=longhorn Microsoft sites say Vista has replaced longhorn. 71.28.243.246 19:58, 13 October 2005 (UTC)
Vista has not been released yet (as of
21 February
2006) and is expected around the end of 2006 (although it could slip into 2007 - my opinion). The article has now been reverted the to reflect this again. Microsoft has info about the Vista beta
[2]
Imagine_B
10:43, 21 February 2006 (UTC)
I reverted an anonymous edit that changed disk to disc. Looking at the wiktionary definition of disk, disc is the North American variant of disk; using this description it is best to keep the terminology consistent. There are roughly three occurances of disk so changing one of them to disc is not a good idea.
The other way to look at it, is in international English (according to wiktionary definition of disc) disc implies "optical disk", in which case hard drive space does not qualify to use disc. BigNate37 19:58, 26 June 2006 (UTC)
Is it necessary to make the comparisons between UNIX and Windows NT? I realize that there is a great schism between the masses of Windows NT and UNIX users, but I don't feel that it would be appropriate to make the claims of:
I'd like to point out that, up to Windows NT 5.0, there *is* a POSIX subsystem (posix.exe and pax.exe) deployed on a standard installation; thus, Windows NT 5.0 and earlier had built-in support for a UNIX platform. This doesn't really fit the description of "features comparable to UNIX," as it contains a UNIX platform. Windows NT's "subsystem" architecture was designed in a manner similar to the idea of "universes" on a UNIX system--it allowed a user to run programs in multiple platforms, including the Win32 subsystem (and 16-bit DOS/Win16 subsystem), the POSIX subsystem, and the OS2 subsystem.
In later versions of Windows NT (5.1, 5.2), the support for the POSIX subsystem was dropped; however, an add-on package from Interix (and now Microsoft) called "Windows Services for Unix" allows for the use of the POSIX subsystem on Windows NT 5.1 and Windows NT 5.2.
Although Windows NT 5.0 and earlier had support for the POSIX subsystem in a standard installation, it lacked the tools that went along with a standard UNIX install. (It's like having a Linux kernel, but no GNU!) The rest of the runtime environment was left in the Windows Services for Unix package.
I believe the second quote (above) is incorrect. SFU is "a POSIX.1-conforming system" [3]; thus, it has source compatibility (up to the POSIX.1 specification, see POSIX) with "UNIX."
Thoughts? Jdstroy 04:36, 4 July 2006 (UTC)
Its purpose was to have a complete implementation of Posix that was absolutely no use to code against, which was achieved by not offering any network API, no graphics API, or anything other than a command line. Why do it? So people purchasing NT boxes on government deals could check off the "Posix compliant" box, which was often a requirement.
SteveLoughran 21:24, 9 January 2007 (UTC)
It may be ineresting to note that MS bundle a number of unix utilities with windows. In the windows release notes its got the disclaimer (the one BSD type licences require you to have). I read that utilities like FTP were rebuilt by MS from sources from the OpenBSD project. —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:47, 25 April 2009 (UTC)
Since increases in processor performance, Windows XP and Windows Server 2003 are actually the latest versions of Windows NT, though they are not branded as such for marketing purposes.
How does the fact that Windows XP and Windows Server 2003 are not branded as NT follow from the fact that CPUs are now faster?
"Released versions of NT for Alpha were 32-bit only"
This doesn't really make sense and I don't think it is even possible. The Alpha is purely a 64-bit architecture. -- Afed 16:55, 18 August 2006 (UTC)
Indeed. The shipped versions of NT for Alpha were pure 32-bit operating systems and supported only 32-bit applications. Same for the first versions of VMS for Alpha, for that matter.
It's true that the chip has no 32-bit mode. This is handled through what amounts to sign extension: Any memory reference on an Alpha has to be done by loading the pointer into a register. And any code compiled for 32-bit Alpha, when loading a pointer into a register, treats it as a signed number, copying bit 31 of the pointer into bits 32-63! Thus the 32-bit OSs on the Alpha populate the very first 2 GiB and the very last 2 Gib of the 64-bit address space, in a manner very similar to the 48-bit "canonical" virtual addresses on X64.
However, unlike 32-bit mode on e.g. X64, all 64 bits of the Alpha's GPRs (and all of the GPRs, for that matter) were available for "large integer" calculations in this "32 bit" environment!
MS released 32-bit NT for Alpha through NT version 4, and shipped some 32-bit Alpha beta versions of Win2K. This continued after CPQ dropped support on NT for Alpha, but with no intention on MS's part of shipping the code. It was done for two reasons: 1) There was Alpha hardware but no Itanium hardware; and 2) they wanted to maintain a code base for more than one processor architecture, a practice that made the (much later) X64 port much easier. But no 64-bit NT code was ever shipped, even in beta form. (I assure you that if it had been, I'd be running a copy!) I imagine that now that they have three architectures to support the Alpha branches have not seen further development. (the above shamelessly plagiarized from my own text on the talk page of Architecture of Windows NT ) Jeh ( talk) 03:23, 14 January 2008 (UTC)
The article says:
What does "to complement workstation versions of Windows that were based on MS-DOS until 2001" mean here? I'm not sure I'd call the DOS-based versions of Windows "workstation versions" - NT was better suited to " workstations" than the DOS-based versions. (The workstation article speaks of UNIX, but that's arguably an error; machines used as scientific and engineering workstations often run NT.)
I'd be inclined to say that NT was originally mainly a workstation and server version of Windows, complementing the DOS-based lower-end consumer and business desktop versions of Windows. It's now replaced the DOS-based versions in the latter market. Guy Harris 00:05, 19 August 2006 (UTC)
Aha, so 'NT' is expanded to 'New Technology' for marketing purposes. And also for marketing purposes Windows 2000 had the slogan 'Built on NT technology.' So this expands to 'Built on New Technology technology.' This must be a joke, right? iNic 02:40, 23 November 2006 (UTC)
I do expect marketing to make logical sense. If the marketing doesn't make logical sense why would their products make logical sense? For example, marketing for New Age products doesn't make logical sense—but neither do their products. So that is fine with me. But maybe MS products doesn't make logical sense either? In that case I will rest my case. (BTW, I've never heard anything but PIN code, never PIN number.) iNic 02:25, 25 November 2006 (UTC)
NT is not about New Technology, but about a HAL9000 like joke... If you take Windows NT, remove the "indows" you will have WNT, and if you do a ROT-1 you will get "VMS" (it's much like "HAL" in "2001: A Space Odyssey", if you to a ROT1 you will get "IBM") - Godzil -- 212.157.49.189 14:45, 26 September 2007 (UTC)
The NT repetitive thing is outlined in this article for any of you out there who want to read about it. Emprovision ( talk) 19:12, 27 May 2008 (UTC)
Like Unix, NT was written in C, a high level language. It can be recompiled to run on other processor systems, at the expense of larger and slower code.
Two problems with this statement: first, what's the basis of the comparison? "Other", "larger", "slower" compared to... Intel x86? Intel i860? MIPS R4000?
More important: A priori there's no reason to believe the claim that recompiling on another plaform gives larger and slower code compared to the original. Without a reference and clarification, this statement should be removed. Bhudson 21:02, 6 January 2007 (UTC)
-If i remember rightly, when NT war originally made (nt 3.x) the dev machines used were deliberately not intel x86 machines. The reason behind this was to stop the developers using nonportable hacks. —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:51, 25 April 2009 (UTC)
Alistair: WinME was a bastard operating system that was released in mid-1999, because MS could not get NT5 out the door in time for the Christmas 1999 season. NT5, rechristened Windows 2000 at the last minute by the marketroids, was released in a big coast-to-coast bash featuring Carlos Santana in mid-February 2000.
By the way, I was a beta tester for both the x86 and AlphaNT versions of Win2k... In fact, I still run RC2 (build 2128) on my DEC AlphaPC 164SX. Dan Schwartz Discpad 23:22, 13 March 2007 (UTC)
Yeah MS wanted to combine the home (DOS) and business (NT) lines into one (like they did eventually with xp) in tje late 90s, but Windows Neptune (codename of the planned OS) was not ready for home use just yet, so they released win2k and ME as a stopgap (me was made in like 6 months).
Thats why XP came out only like 9 months after windows 2000, and they are very similar. Windows XP was basically windows 2000 + the home stuff like the application compatibility wizard, network setup wizard etc.... planned for neptune.
If you try the windows neptune alpha (some on the net) you will see what i mean. Its basically a very buggy win2k, with some wizards that didnt emerge until xp (like the beginnings of a welcome screen implementation) —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:54, 25 April 2009 (UTC)
The article says Home Server will be NT 5.2. If there is a source to this claim please post here if not I will change it to:
NT Ver. | Marketing Name | Editions | Release Date | RTM Build |
---|---|---|---|---|
Unknown | Windows Home Server | Unknown | 2007 (expected) | Unknown |
Chetblong 20:03, 1 June 2007 (UTC)
I went ahead and changed it to Unknown. -- :) Chetblong 19:02, 7 June 2007 (UTC)
It is 5.2. I own it. Its a version of 2003 Small Business Server with domain functionality and IIS removed, among other things + addons like the new volume manager and backup.
The file version it comes up as (on nlite etc...) is 5.2
[all nt <vista can be installed to fat..] "at the expense of speed and security" no, fat is faster than NTFS for normal use on all but the very largest drives.~blab
;Only on small partitions running DOS! Per KB: "over 200 MB the FAT file system should not be used. This is because as the size of the volume increases, performance with FAT will quickly decrease." per http://support.microsoft.com/kb/100108 . Also consider problems like more than a few thousand entries in a directory (like Internet Cache).
"Limitations of FAT32 File System". Microsoft Help and Support. 2004-12-16. http://support.microsoft.com/kb/184006/en-us Chen, Raymond (2006). Microsoft TechNet: A Brief and Incomplete History of FAT32. TechNet Magazine July 2006 http://groups.google.com/group/microsoft.public.win98.fat32/browse_thread/thread/1fa5783ca7a26a5b
Shjacks45 ( talk) 13:20, 13 April 2010 (UTC)
The Hardware requirements section doesn't mention year difrences between the respective operating systems, something I feel may make it be unbiased against Microsoft for the percieved software bloat (noted, they rely less on application optimization and more on Moore's Law in order to get speed increases). If the years were listed in the table, perhaps it would be more effective at showing the data? - 68.228.56.158 19:13, 20 July 2007 (UTC)
Which is the most stable NT version? I know the kernel is refined all along the way, and NT 5 (Windows 2000) included a major architecture redesign. However, what about the whole system's buglessness and stability? In Yahoo! answers one said it was Windows 2000 that was the most stable version. I myself felt that too (patched at least up to year 2004). In my experience XP is also stable, however with some strange user mode hang-up situations: when switching users, logging off, etc. -- 211.167.159.70 08:06, 14 August 2007 (UTC)
-the server builds are the most stable IMHO. Server 2000 or 2003 have never crashed for me under general use, unlike XP and 2000 Workstation —Preceding unsigned comment added by 86.16.160.17 ( talk) 21:45, 2 January 2008 (UTC)
Depends on what you consider stable. Downtime? In which case probably Server 2008 - its got self-healing NTFS and hotpatching. Windows 2000 is mature - but that doesnt nececerially mean stable. —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:58, 25 April 2009 (UTC)
Does the controversy concerning the minute different versions of NT Server (4.0) and NT Workstation (4.0) deserve to be brought up in this article? http://www.oreilly.com/news/differences_nt.html —Preceding unsigned comment added by 75.74.240.92 ( talk) 02:56, 27 November 2007 (UTC)
Release list should include Home Server and processor type for all releases. Hel pslo ose 21:08, 20 December 2007 (UTC)
NT was written in C, a mid-level language. This means that it can be compiled to run on several processor systems; however, the code produced by the compiler is larger and slower than assembler code written for a particular processor[citation needed]. For this reason, NT was not favored initially for use with slower processors with less memory.
As a CS engineering student, I can almost certainly say this is wrong. Code produced by a compiler is not inherently slower, and it will CERTAINLY not be the reason why a certain OS cannot be used for a certain machine. Heck, Linux avoids ASM code where slightly possible, only very lowlevel function that cannot be described in C will be programmed in ASM. —Preceding unsigned comment added by 193.190.253.149 ( talk) 15:52, 11 January 2008 (UTC)
What it means is that NT is slower than 9x.
9x had lots and lots of asm code, which was highly optimesed for x86. —Preceding unsigned comment added by 134.36.92.18 ( talk) 23:59, 25 April 2009 (UTC)
What do you think of the table now? I confess I'd forgotten completely about XP 64-bit Ver 2003! (and MS did too, very soon after...) I have to run out now, but later I'll put in a link to that press release.
yeah, there is XP x64 (Intel Itanium architechtute) and XP Pro x64 (the common x86_64)
XP x64 (itanium) was pure xp SP1. XP Pro x64 was based on Server 2003 x64
Do you think the supported processor architectures should perhaps have a column of their own? This might mean a few more new rows, but it's arguably useful to have it in the table. Jeh ( talk) 03:15, 14 January 2008 (UTC)
XP 64 bit (Intel Itanium) was based on pure xp, not 2003. It was only supported up to SP1 i think. XP ~Pro~ x64 is supported up to SP2 and is based on Server 2003. Dont get them confused.
134.36.92.18 ( talk) 00:12, 26 April 2009 (UTC)
Do we even have a source for this? If not, it should be removed. I doubt Windows 7 will be NT 6.1, especially considering it's Windows 7, which comes after Windows NT 6/Vista. 99.149.127.80 ( talk) 06:43, 5 April 2008 (UTC)
MS said (in a webcast on msdn discussing windows driver development for windows 7 for OEMs) that windows 7 will be NT 6.1, as will windows server 2008 R2. This is in order to keep compatibility apparently. —Preceding unsigned comment added by 134.36.92.18 ( talk) 00:01, 26 April 2009 (UTC)
I doubt it also that it will be NT 6.1, it's not logical, and further more my I heard that Microsoft named 'Windows 7' Windows 7 because they want to go back to the naming of their OS with the version number... Like they did in the beginning... Windows NT 3.1, Windows NT 3.5, Windows NT 3.51, etc... Wouser ( talk) 13:08, 18 May 2009 (UTC)
It is DEFINATELY 6.1 thats what it says in the windows 7 release candidate and what i read in some prelease material aimed at OEM driver developers. So will be Server 2008 R2 (which will be based on windows 7, as opposed to the original server 2008 which was vista sp1). Windows '7' is purely a marketing term. it is NOT nt 7. —Preceding unsigned comment added by 134.36.93.46 ( talk) 03:07, 21 May 2009 (UTC)
There's precedent for the official marketed version number and the actual OS-reported version number being different: Windows for Workgroups 3.11 actually reports its version as "3.10". --20:47, 4 May 2011 (UTC) —Preceding unsigned comment added by 69.116.117.48 ( talk)
Windows Server developer Preview uses 6.2 Kernel Winver. — Preceding unsigned comment added by 79.237.254.50 ( talk) 10:56, 8 February 2012 (UTC)
I think this source is certainly more reliable than this one. Accordingly, if no one has any objections, the Xbox line should be removed. Or at least reworded so that it says the Xbox OS is derived from the Win32 APIs but built from the ground up. - xpclient Talk 11:41, 22 May 2008 (UTC)
Windows 7 is still a codename and not a "marketing name", Windows 7 is schedueld to be released in H1 2010, not H2 2009. —Preceding unsigned comment added by 129.241.150.119 ( talk) 15:55, 30 August 2008 (UTC)
Windows 7 WILL be sold as "Windows 7". Its the release/marketing name now, not just a codename. —Preceding unsigned comment added by 134.36.93.46 ( talk) 07:13, 22 May 2009 (UTC)
See this. I don't feel like creating an account to fix this/cite that part of the article. Check the second paragraph. —Preceding unsigned comment added by 69.84.252.227 ( talk) 00:28, 31 January 2009 (UTC)
There's some disagreement regarding whether the TCP/IP implementation is derived or imitating BSD. However, Winsock is derived from BSD. Tedickey ( talk) 14:06, 4 April 2010 (UTC)
see xbox, dreamcast I worked at Microsoft in 1997 when NT 5.0 beta code was forked to Xbox. No Win32 or other subsystems. Windows Native Mode, Drivers, and DirectX; all kernel mode stuff. Not sure of Xbox 360, but there was PowerPC NT 5.0 code available at the time (except for Internet Explorer). Dreamcast was WinCE with DirectX, bare bones CE will fit in "250K" but not directx. Most modern BIOS is 128KB (to 512KB+) which includes chipset setup and EFI boot code then shrinks to 64KB after boot. Fact is most games load their operating system from the game CD or game cartridge, see Playstation "disk swap" section. There was a project there to "secure boot" PCs by putting MBR code and boot files into ROM (Flash) to prevent "viruses". NTLoader, NTDetect, boot.ini et al can fit into 250K. PicoXP is only 14MB. Shjacks45 ( talk) 14:48, 13 April 2010 (UTC)
"Initial development was for the Intel i860 RISC then MIPS until i386 became available.[5][6] ". According to the wikipedia articles about these processors the i860 was produced from 1989 and the i386 from 1985. The references 5 and 6 are mostly circular and are not useful. —Preceding unsigned comment added by L00KnS33 ( talk • contribs) 07:48, 11 August 2010 (UTC)
Windows 2000 64-bit was codenamed "Janus".
So how come it's not on Windows NT#64-bit? Also see Talk:Windows 2000#64-bit. - 79.179.200.231 ( talk) 19:34, 9 January 2011 (UTC)
Here's an old Microsoft page about it:
WhisperToMe ( talk) 01:58, 14 October 2011 (UTC)
Steve Kleynhans said: "Windows 8 is more than a major upgrade to Windows - it's a technology shift. We don't see technology shifts too often; the only other one Microsoft's client OS has gone through was the move from DOS technology to Windows NT technology, which began in 1993 and took eight years, ending with the introduction of Windows XP in 2001,"
News link: http://www.ibtimes.com/articles/356436/20120626/windows-8-marks-beginning-winrt-winnt-gartner.htm — Preceding unsigned comment added by 77.121.210.102 ( talk) 15:38, 3 July 2012 (UTC)
This also uses an NT 6.x kernel. Should it be mentioned? PizzaMan ( talk) 12:59, 26 May 2013 (UTC)
The Page is Currently Locked and Contains Information that is Uncited and I Know is incorrect. Windows 3.1 Didn't Run on the NT Kernel it ran on the DOS kernel and I know this for a fact. Someone Should Correct this I think it's more than a simple mistake.
The quote as given is not suitable for the technical description, since the point of it was to accuse competitors of being dishonest. There's no technical issue presented in the comment. TEDickey ( talk) 13:16, 13 October 2013 (UTC)
I would like to see a citation supporting the following claim:
"Windows NT is a family of operating systems produced by Microsoft, the first version of which was released in July 1993 ... Windows 2000 and its successors are members of the Windows NT family, although they are not branded using the name 'Windows NT'." and for the inclusion of Windows 8.1.
I don't see any justification for calling every version of Windows until the end of time "Windows NT". I would have less of a problem if we differentiated between Windows NT the operating system and the Windows NT family of operating systems and if we used Category:Windows NT family of operating systems or even Category:Windows NT architecture instead of Category:Windows NT, but most of the references I can find for "Windows NT family of operating systems" are rather old and in the context of explaining how this new OS is different from the Windows 95 family of operating systems. I would very much like a citation to a reliable source that tells us which operating systems are or are not part of the Windows NT family of operating systems. -- Guy Macon ( talk) 22:16, 22 April 2014 (UTC)
Guy Macon, Jeh, Dogmaticeclectic, Guy Harris, Codename Lisa, I decided to tackle the issue of finding a WP:RS that ties all of this into a "family". Microsoft's OSINFO structure's article says:
The use of the word "descended" is still vague on if we can call this a "family" but is getting closer. Microsoft's OS_TYPE Enumeration links the family a bit better in that it classifies machines as one of:
Click on <224> above to learn that a "Windows NT workstation" indicates "Windows 2000 Professional, Windows XP, Windows Vista, Windows 7, Windows 8, or Windows 8.1."
Click on <226> above to learn that a "Windows NT server" indicates "Windows 2000 Server, Windows Server 2003, Windows Server 2008, Windows Server 2008 R2, Windows Server 2012, or Windows Server 2012 R2."
I had also looked at the WIN32 API's GetVersionEx() function which returns an OSVERSIONINFO structure that contains the dwPlatformId field. This field is defined as holding the value 2 or VER_PLATFORM_WIN32_NT meaning "The operating system is Windows 7, Windows Server 2008, Windows Vista, Windows Server 2003, Windows XP, or Windows 2000." While that Microsoft article talks about Windows 8 and 8.1 it's silent on what the dwPlatformId is for those operating systems. I've submitted feedback to Microsoft to see if they will fix that detail.
I suspect Microsoft's OS_TYPE article and its footnotes are decent WP:RS that Windows NT and its descendants up through Windows 8.1 / Windows Server 2012 R2 are in the same family. I'm not sure how to work that into the Wikipedia article without a bunch of hand waving. I'd like to "show" rather than "tell" the WP reader it's a family.
In my own mind I've always called it the "Windows NT architecture" rather than a "family" but in reading over this thread it seems "family" is a better word and that we use "Windows NT" to separate this from the older generations of the Microsoft Windows family. -- Marc Kupper| talk 22:20, 11 September 2014 (UTC)
The result of the move request was: not moved. -- BrownHairedGirl (talk) • ( contribs) 17:54, 8 May 2014 (UTC)
Windows NT →
Windows NT family (typo corrected; was "Windows NT Family") – This rename will differentiate between Windows NT the specific operating systems named "Windows NT" and Windows NT the family of related operating systems that includes Windows XP, Windows 7, and Windows 8. --
Guy Macon (
talk)
20:55, 26 April 2014 (UTC) Edited 12:04, 27 April 2014 (UTC)
*'''Support'''
or *'''Oppose'''
, then sign your comment with ~~~~
. Since
polling is not a substitute for discussion, please explain your reasons, taking into account
Wikipedia's policy on article titles.Recently all versions of Windows (XP+) have been changed to Windows N.T. ¿Should Windows Phone 8 and later also have this on the top of their infoboxes? as they are also based on the Windows N.T. kernel ¿or are they different because of other reasons? Sincerely, -- Namlong618 ( talk) 16:45, 18 February 2015 (UTC)
The Windows NT family includes mobile, embedded, desktop/notebook, and server OSes. The infobox's release information is for the desktop/notebook flavor; should there be a separate infobox for the server flavor, and perhaps also on the Microsoft Windows page, or should that go on the Windows Server page? This page doesn't give a marketing target in the infobox; the Microsoft Windows page does, giving it as "personal computing", but it also covers non-desktop/notebook flavors of Windows. Guy Harris ( talk) 23:20, 19 July 2015 (UTC)
Current version says that the 10-fold disk space increase in Vista is primarily due to the extra memory needed by the Desktop Window Manager. This doesn't sound right -- it's mixing system memory / VRAM with disk space. The Vista DWM was notorious for using a lot of memory and causing slow performance as a result -- which was fixed by rearchitecting it in Windows 7 -- but I've never heard of it being a significant cause of disk space usage and no citation has been provided for this claim. 24.130.133.67 ( talk) 09:18, 22 July 2015 (UTC)
Windows 10 is actually 6.4 despite the 2014 news stating that it will be 10.0. Someone should clean up the article. ebertek ( talk) 15:53, 24 March 2016 (UTC)
Shouldn't there be a mention that NT was bought from Northern Telecom? — Preceding unsigned comment added by 67.78.5.254 ( talk) 15:45, 8 April 2016 (UTC)
There is currently a discussion of whether Template:Infobox OS should be used with multiple version numbers - for example, to list both a "software update" and "next major release" beta, or to list betas from more than one release stream. If you believe that multiple {stable, preview} releases should never appear in that infobox, or if you believe that they should appear under some or all circumstances where there's more than one beta of the OS in question available, you might want to comment there. (I have no strong belief either way; I'm OK with the main OS page listing only the "next major release" beta, but listing betas from multiple streams if they exist, but I'd also be OK with other choices.) Guy Harris ( talk) 08:09, 11 July 2016 (UTC)
Please see Talk:Windows NT 4.0 § NT 4 kernel source code on GitHub — Codename Lisa ( talk) 07:07, 21 May 2017 (UTC)
Not that I want move this page to a new title, but is there any reason it is still move protected over a single instance of page move vandalism from 8 years ago? It doesn't seem to meet the purposes set out in WP:MOVP. Kb.au ( talk) 14:42, 30 June 2017 (UTC)
This article is such a page. In applying all Wikipedia policies, common sense is a requirement. In this case, the decision to protect it is just good sense.Highly visible pages that have no reason to be moved
I think the following paragraph (last in the overview section) uses encyclopedic language, and could be unnecessarily promotional, but I didn't want to make changes since I don't often visit this page and am not familiar with the history:
"In some ways the ancestors of Windows NT are the operating systems that Dave Cutler worked on before he was hired by Microsoft, namely the VMS and RSX-11 operating systems, and also an unreleased object-based operating system developed by him for DEC Prism. This lineage is made clear in Cutler's foreword to "Inside Windows NT" by Helen Custer.[4]"
The phrase "in some ways" is clearly nonspecific and those "ways" should probably be enumerated. Also, I think the general tone of this sounds a little promotional towards Dave Cutler (he is the author of the source cited to prove the claim that he developed the "ancestors" of Windows NT. If he is really a big figure in Windows NT that's obviously fine, but this sentence doesn't make clear what his contributions were, or why they were important enough that he is listed in the overview section rather than the history section.
Basically, when reading this section I felt kind of like I was talking to some weathered Computer Scientist giving me his version of how Windows came to be, not like I was reading an objective encyclopedia. — Preceding unsigned comment added by Ludicous ( talk • contribs) 16:45, 8 September 2017 (UTC)
Why the new Windows logo is being used for this 1993 released OS? BarricadeX ( talk) 04:00, 14 August 2018 (UTC)
This page should not be speedy deleted as pure vandalism or a blatant hoax, because I WANNA DELETE THIS PAGE — Preceding unsigned comment added by Ziadelwan ( talk • contribs) 19:08, 29 August 2018 (UTC)
The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for deletion:
Participate in the deletion discussion at the nomination page. — Community Tech bot ( talk) 10:40, 30 June 2021 (UTC)
2401:D800:2220:8C36:3152:2512:9718:BA30 ( talk · contribs · WHOIS) / 2401:d800:2de0:e28e:56fd:f712:2d14:f7fa ( talk · contribs · WHOIS) (I'm assuming they're the same person) has recently tagged this page – and hybrid kernel – with {{ Citation needed}}, stating that we need official confirmation from Microsoft to call NT a hybrid kernel. I don't see how that is grounded in policy, though. WP:V requires reliable sources, so if reliable sources call NT a hybrid kernel, that should be stated. I don't know of any policy that would additionally require confirmation by the vendor of a product in a case like this. ;; Maddy ♥︎(they/she)♥︎ :: talk 08:30, 24 August 2022 (UTC)
The idea behind a hybrid kernel is to have a kernel structure similar to that of a microkernel, but to implement that structure in the manner of a monolithic kernel. In contrast to a microkernel, all (or nearly all) operating system services in a hybrid kernel are still in kernel space. There are none of the reliability benefits of having services in user space, as with a microkernel. However, just as with an ordinary monolithic kernel, there is none of the performance overhead for message passing and context switching between kernel and user mode that normally comes with a microkernel.
}The Windows NT 4.0 architecture merges the best attributes of a layered operating system with those of a client/server or microkernel operating system. [2] The Windows NT Executive is the generic name for a number of subsystems and components of the operating system that execute in the microprocessor's kernel (or privileged*)* mode. The remainder of the operating system, such as various server subsystems and their application clients, execute in user (or application) mode.
A microkernel, on the other hand, is the name given to the core portion of a modern, modular operating system. Microkernel operating systems are based on two fundamental principles. The most basic principle is one of modularity, encapsulation, and data hiding. In this aspect of the design, there is one and only one portion of the operating system that has system-wide responsibility for a particular function. All other parts of the operating system (as well as applications, naturally) access that function through a well-defined interface. There is no duplication of function and no "back doors" to critical data structures; all access to system data structures is through software interfaces. This approach makes it possible to upgrade or replace entire modules within the system without disturbing the remainder. A secondary principal of the microkernel design, related to the first but focused more on the implementation strategy, is that large portions of the operating system which traditionally run entirely in the kernel or privileged mode of the microprocessor can now be executed in user or application mode, with only the microkernel itself, along with a relatively small amount of hardware device-specific code, executing in kernel mode.
CreateFile()
sending a message to CSRSS, which then constructs a path name that it then passes to the NtCreateFile()
system call, the result of which it sends back.)open()
is a system call, unlike CreateFile()
, and fork()
, the exec
calls, and posix_spawn()
are system calls, unlike CreateProcess()
. I'm not sure how its graphics/window system layer currently looks, but I suspect it may not be too different from current Linux systems using
Wayland. With DriverKit, some drivers - including, from my experience with Ethernet driver crashes when running
macOS Ventura on a
VMware Fusion virtual machine, at least one Ethernet driver - may have been moved from the kernel to user mode processes.References
As to the whole "hybrid kernel" thing - it's just marketing. It's "Oh, those microkernels had good PR, how can we try to get good PR for our working kernel? Oh, I know, let's use a cool name and try to imply that it has all the PR advantages that that other system has.