He is also on our forum - https://forums.mydigitallife.net/threads/convert-windows-server-to-workstation-offline.84632/
They still recycle the same info they published in 2003, which were valuable, but now are incomplete. No word about the missing audio codecs, for example (which is a thing only since server 2016) https://forums.mydigitallife.net/th...ible-alternatives.61061/page-529#post-1789190 No word about the power plans, which are way different between server and clients (just export /delete /import HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Power) to go from server to client or viceversa. And so on...
I consider the knowledge about Windows firstly an hobby. If you start to bureaucrati-fy it becomes a job. And fine to share my hobby for free, but I want to be paid for my job.
Thats true, you're the expert For a beginner like myself, I need those shortcuts. I'm not on your level yet, but I have faith I"ll get there.
This, is really cool. But if you 'stay pro' by replacing the branding folder, as you describe -- what exactly do you have here as far as the OS underneath, and what happens after you hit Windows Update?
You have what I explained in previous post. The Pro (or other client SKUs) policies. Which means the ability to install many client SW an drivers that are artificially blocked on Servers Nothing special, WU will follow the original SKU (on a server converted to client server updates will be offered, on a client converted to server client packages will be offered)
I'm sorry, I should have been clearer: iirc Server used to run a different ntoskrnl.exe, and other system files than the desktop OS... If this is still the case what happens in the future with "SKUwitch" as far as Windows Update could get interesting... I'm sure you must have some idea how much Microsoft likes user tinkering with their Server OS...
When? Surely 2003 used a newer kernel than XP32. But that was just a different generation, not a purposely differentiation between client and servers. Indeed XP64 (which is younger than XP32) uses the same kernel as 2003 x64. Differentiation is made by kernel policies, installed packages and few custom registry settings. If MS cared about that (speaking of individuals/hobbists/entusiasts) they had blocked kms emulators, and sites like MDL would have been taken down in zero time. In short, yes I have a precise idea about how much they care. What they officially call piracy, unofficially is what made windows the dominant OS. Now that they started to sell blades instead of razors, they care even less than they used to.
Yes, that dates my apparently mistaken assumption almost exactly. Circa Server 2003 R2 Mark Rusionvich had joined the company, was a very compelling spokesperson for Windows System Internals, gave a cool recitation and wrote several papers on future Windows Server kernel development that made it at least sound like there would be kernel developments exclusive to the Server platform... Looking back this is also about the same time I think David Cutler was stepping down from leading OS design and development, and there was a more concerted effort to move more code out of the kernel to services and processes... Kinda sad what happened since -- as there were some really cool and aspirational design intentions for the Windows Operating Systems. Not saying there aren't talented and aspirational people working for Microsoft today; I think Cutler is still there in his 80's in some capacity -- but the bulwark of bulls**t looms epically larger than ever two decades hence...
It's easier to develop one kernel then block and/or enable some features than develop two different kernels. That said the windows kernel has improved vastly over the last decade, just think to native vhds, wimboot, Hyper-V, RemoteFX, deduplication, WSL and WSLg, GPU P..., most of them have no real replacement in Linux, let alone the useless MacOS It's just that MS (for unknown reasons) likes to advertise loudly pointless "improvements", like rounded corners over square ones, or new emojis, while whispering or staying completely silent about real innovations. About the graphic drivers moved mostly outside the kernels to get stability over performance, they just changed their minds multiple times. WinNT 3.x started with user mode drivers, in NT4 they switched to kernel mode ones, in vista they stepped back to user mode ones, then generations of DirectX blurred the lines, there isn't a right way to do it, there are just different compromises and tradeoffs.
While Microsoft may have 'improved' the Windows kernel, invented, branded, registered, (and/or even patented & licensed) some of these acronyms -- the concepts of what they describe and portray; booting an OS image file that runs native drivers from an image 'on bare metal', virtualization, OS emulation, remote display protocols, remote virtualization, GPU/CPU (and other hardware) virtualization/paravirtualization and/or pass-through were available on Linux and UNIX long before Windows, in most cases by more than two decades... As well, in 'don't know how to code Userland', with UNIX and Linux you've never have to wait for Microsoft to fix a kernel issue, limitation, add or remove a feature, better optimizes it for your use case, issue, or hardware -- you can do it yourself, can custom compile a kernel with just the features you want (and none you don't) and optimize the kernel for exactly the hardware you are running -- and have been able to do this since before Windows had a kernel again with no exposure to coding with distros like Gentoo Linux, Open BSD UNIX, and many others... Heck SLS Linux was offering this and RDP virtualization in 1992...
Please.... don't start with premade pointless open source propaganda, I use Linux since the days of Redhat 4, and I know what are the strength and the weakness of Linux. But I like to be honest. RemoteFX (now dropped) and GPU - P have NO equivalent in Linux nor Macos, nor BSD. Period. There is no type 1 Hypervisor that can do some of the things that Hyper-V can do, Period Although installing an OS over a n image file is not an absolute novelty, to get a native vhd lookalike, I spent literally days customizing scripts, initramfs images, kenrels and so on, to get a system that will be destroyed by any major update, being non standard. As for deduplication while ZFS (which was developed by a company, not a community) can do that, but requires 8GB of ram per TB of storage, while I use windows' deduplication on 2TB of storage Using a server that runs on 8GB, but runs as well another three virtual machines, something that is possible thanks to the dynamic ram provided by Hyper-V, another thing that has no equivalent in KVM, Xen or whatever. I can continue for a day. But back to your point that MS kernels hasn't evolved, I must say it evolved more and better than Linux one. But we get the systemd cancer instead. Sure you can do some of the thing you mention, if you're skilled enough, and if you have a ton of time to spend (or waste). Try to run a modern Linux on my Acer 5104 from the XP era, while still having a working 3D and most of the features like suspension, hibernation. For fun I did it, and I spent a week recompiling kernels, downgrading Xorg components, mixing packages and so on. The very same notebook runs the latest Win10 or Win11 w/o a blink using drivers last updated 15 years ago, w/o a single manual tweak. That's windows. We can debate about privacy things or pointless GUI regressions, but absolutely there is nothing to debate about its kernel and basic structure, they are the best in class, unless we talk of niche applications, and they had an incredible evolution, a remarkable evolution that mostly didn't break even 25 years old things.
Oh, so you're the final authority on Linux, UNIX, and BSD... I had now way of knowing... It would be better to actually be honest, and informed, rather than just 'like to be honest' -- I recommend it, and making more of an effort in that regard rather than 'pointless' blowing your horn, and shilling commercial products... But you're obviously The Final Authority and above reproach... Many would argue it takes far less skill, and time, than having to resort to non-existent Microsoft documentation, and forums with deliberately obtuse and obnoxious people that suggest they 'don't have time' to explain their 'hobby', non-sequiturs and pidgin English documentation of undocumented Windows System Internals...
I'm just a person that in the last 42 years has used and has seen the evolution of practically any OS (including a small bit of contribution to the linux kernel) Then free to agree or disagree. But if you disagree is better if you do about something you know, not about something you heard. So if you are able to write such sentence what prevent you to read it and act consequently? Moving the target of a discussion never helps. You stated that the windows kernel didn't evolve and argued that's a false assumption, based on said rather than on facts. Nothing less nothing more.