Microsoft fixes Point of Sale bug that delayed Windows 11 startup for 40 minutes A fresh Windows 11 patch slipped out overnight as an optional update, but contains an impressively long list of fixes for Microsoft's flagship operating system. One bug addressed in KB5012643 could leave Point of Sale terminals hanging for up to 40 minutes during startup. The content of the release, 22000.652, had previously …
>Yes, it is 2022 and Windows can still leak memory if you don't turn it off regularly.
"Yes, its 2022 and Windows still leaks memory" is what you meant to say. Turning it off doesn't stop the leak.... no, wait, it does if you don't turn it back on again.
/mines the one with Pedant stencilled across the back.
Our vets is still running VT terminals into their backend. And it works amazingly well still. I've lived here 31 years and been going to the same vet for 30 of those years. The actual terminals have changed - most of them are now LCD screens running Windows with a terminal emulator, so they can do email as well, but there are still some of the consult rooms with VT 100s.
Airline staff still prefer their mainframe menus for ticketing. The initial learning time may be longer, but it is massively more efficient that the various GUIs companies have tried to place on top of it.
I personally still prefer a command line - being forced to use a mouse is not only irritating, it seriously slows me down..
SAP has Mostly replaced our VMS based system that used to manage the whole business from Warehouse to Finance on two servers in a cluster. SAP and other systems that mainly do the finance only, now uses 150 servers and the warehouse stuff will add another 30 odd.
Sometimes I wonder if all the pretty icons are worth it.
VMS also is totally secure by obscurity, though I believe Sophos does or at least did at one time do an Anti Virus solution for it.
You beat me to it. Why in Heaven's Name would anyone be risking upgrading to run POS (or any such critical system) on a new operating system like this?
I have allowed it to be installed on my non-critical Surface laptop. My immediate observations are: a) it takes longer to boot (not a lot, but noticeable), and b) every single time you boot up there are updates, requiring further reboots (I don't use the Surface every day, just occasionally, and that is probably why). Some features - the mouse as one example - can stop working until you reboot. Other than that, it is working fine as I am evaluating it.
There is no way it is coming near my PC until I can be sure all my software works properly on it. Eventually, of course, it will have to.
A client has a Surface which he uses for business critical work. In particular, he has bespoke client-centred 3D design software on it, and that was problematic enough in its system requirements under W10 (it uses a dongle for security). It is that grey kind of software that is advanced (and expensive) and yet totally non-standard in many respects.
He asked me if he should upgrade and I told him absolutely not. No way. Not yet. Because it was almost guaranteed to stop his design package working given the slow rollout of updates by his software vendor and our previous issues getting it to work properly.
My immediate observations are: a) it takes longer to boot (not a lot, but noticeable), and b) every single time you boot up there are updates, requiring further reboots
Time to give LinuxMint a try. Been using 20.3 for several months now without problems (and other versions for the last 12 years). I don't know how my laptop compares with a Surface but it has consistently booted from cold within 6secs from POST to login. MS users complain that hibernate doesn't work on Linux but when it boots quickly enough why use hibernate at all? All updates happen in the background (and only if you accept them) with no noticeable slowdown while you are using the desktop and the next boot will be just as quick because the whole update is completed BEFORE you shutdown - none of that waiting for updating 2 of 3, then 3 of 3.
My client's expensive software barely works on Windows (every iteration has been problematic, and it was designed for use specifically on Windows). It is dependant on Heaven knows what hardware, memory, and graphics requirements even if you meet the main spec (and the last upgrade jumped the spec by a mile, which is why he ended up using a Surface). Once it is going, it does exactly what it is supposed to do reliably. Just not when installing it.
It does not run under Linux or anything else. Just Windows. And it is critical to his business - absolutely critical.
He is not a hacker/programmer/tinkerer. He is a user. His company turnover is based on him being user. So even if the software did run under Linux (and it doesn't), the learning curve for him to run it and use it would be considerable.
Sometimes, Windows really is the best (and only) option for some people. And my client is one of them.
But point taken about the alternative option.
This is the maybe 5th major iteration of NT which was first released as NT 3.1 nearly 30 years ago (1993).
It's really a fiddled with version of Windows 10.
Are they rushing stuff out with little design and less testing? It's an appalling list of fixes.
Note:
NT 3.1. NT 3.5 (possibly 1st Server & Workstation artificial versions), NT 3.51 (add faked for Win95 apis not in NT3.5 or Win32s intended to stop people running Office 95 on Win32s on Win.3.1, Win 3.11 & WFWG3.x).
NT 4.0 (Add Explorer shell and stupidly move Graphics driver into Kernel for 10% performance). Took sp6 to fix all. Sp7 would have added USB but cancelled to save Win2K sales). Note that original Win95 had no USB.
Win2K, XP 2003: really NT 5.x. The third version
Vista, Win 7: really NT 6.x, win7 should have been free to Vista users. The 4th version
Win 8, a sort of Phone GUI version of 7. Should never have existed on desktop.
No Windows 9 because idiots during NT 4 & NT 5 era checked for Win9x
Windows 10. Really 5th version because Windows on Desktop was an aberration. Mostly free to Win7 & 8 users. 4th or 5th iteration.
Windows 11. Increment version number for Marketing reasons. See version numbers of Word For Windows!
"Guess Doug Cutler finally gets the last laugh."
Dave Cutler has been (and afaik still is) a Microsoft employee for many years (since 1988?). He's not long passed his 80th birthday.
https://en.wikipedia.org/wiki/Dave_Cutler
Sadly although VMS on AMD64 is here and now (after a fashion), the company responsible for developing and marketing it seems to have even less of a clue than HPQ.
True, but why was it only in NT4.0 as a technet preview in 2000 approx? USB May 1996, and NT4.0 August 1996. USB was never publicly released for NT4.0, presumably so as to protect Windows 2000 sales. Yet Win2K was rushed out, buggy and unfinished. We continued installing NT 4.0 sp6 for customers until SP1 of XP was out. XP SP1 was the "completed" version of Windows 2000 (NT 5.1 vs 5.0).
Was the Itanium version of XP the shortest supported windows ever?
I did test the technet/MSDN preview of USB on NT4.0. The scanner wouldn't install because the programmer was checking OS Version. MS frequently told people NEVER check for OS version to decide to install but installed features! So I installed driver on a Win2K laptop and copied the files to the NT4.0 install with added MS USB stack. The scanner worked.
It's amazing though how many people assume Win95 did have it from the start.
Baffled why ANY ATM or Till (POS) has used Windows in last 20 years. Linux was certainly stable and suitable for those even in 1999, certainly by the time Vista arrived it was a far better choice. Though when regular users were getting Vista the POS might have been NT 4.0 embedded or Windows CE. Setboxes, routers, Sony ereader all linux then. BT switched existing install base from WinCE to Linux of setboxes.
The ONLY thing absolutely needing Windows was Workstations, Laptops etc. Though we didn't ditch Windows for Linux completely till we didn't want server based Windows Update Service.
>> Baffled why ANY ATM or Till (POS) has used Windows in last 20 years. Linux was certainly stable and suitable for those even in 1999,
Do you have actual experience with POS systems?
Microsoft has for a very long time now (since 90s?) had the OPOS subsystem which allows any point-of-sales device to register itself as a customer display, receipt printer, cash drawer, magnetic stripe reader, barcode reader and so on. The actual till software can easily interact with all of them without having to know the hardware details, whether the device is in serial or USB port, for example.
In 1999 the driver selection for such things was poor or likely just nonexistent. In 2000 I was working in an fast food chain that was changing from Wyse terminals to touch screen POS systems, and I was evaluating candidates from different vendors. Unsurprisingly, no vendor back then provided any support for Linux.
Installing Linux wasn't the hard part since these were mostly x86 computers in a funny shaped box and a standard IDE drive. It was the touch screen that didn't have drivers at all or didn't function well enough. The built-in magnetic stripe reader didn't have drivers. And so on. No such problems with Windows because the POS vendor provided everything. Call the appropriate function or driver and presto! - the cash drawer opens.
(That Delphi codebase was used until I left a couple years ago; only a few years earlier the few surviving Windows 98 based systems were replace due to EMV libraries requiring at least WinXP)
"Microsoft has for a very long time now (since 90s?) had the OPOS subsystem which allows any point-of-sales device to register itself as a customer display, receipt printer, cash drawer, magnetic stripe reader, barcode reader and so on. "
Standards-based computing for device and platform independence, what an excellent concept. Better open standards than properietary, but that's asking a bit much of MS. Meanwhile...
POS for .NET, is that what you meant?
What's it been up to lately?
Maybe this?
https://docs.microsoft.com/en-us/windows/uwp/devices-sensors/pos-get-started (2020?)
Or for device support, maybe this?
https://docs.microsoft.com/en-us/windows/uwp/devices-sensors/pos-device-support (2020?)
Or maybe this?
https://www.suse.com/products/linux-point-of-service/
https://documentation.suse.com/sle-pos/11-SP3/single-html/SLEPOS-guide/#book.slepos
If you'd every worked on POS systems as an implementer you'd know it's the same reason why gaming is overwhelmingly Windows - devices just (mostly) work. And the problem is far, far worse on the POS side even than on the gaming side.
You can have a POS system, the printer goes bad, so you just plug in a new printer - even if it's not the same model Windows finds it and uses it. Adding new printers to a linux system is still a crapshoot. Now add in an incredible range of scanners, scales, cash registers, interfacing with the payment system, etc. etc. Windows handles a lot of this transparently while there aren't even Linux drivers for most of these.
Of course that's a chicken and egg problem, but one that Linux has never really cared to solve (or has tried to in various incompatible ways). And that's why people use Windows on POS. Even if you do have to reboot the damn things every morning, the support load is much less.
Now is there any excuse for running Windows *11* on these things? Not yet.
I'd much rather do that than try to get printing working on linux when it doesn't just work.
More pertinently, the hideousness of Windows Home has very little to do with Windows POS, where things really do mostly just work.
I say this as someone with near equal numbers of Windows and Linux machines and who's dealt with how shitty they can both be for 20 years. It's good to be honest about the plusses and minuses of each. For instance, all my servers are Linux, Windows servers are a f@#$ing nightmare. And similarly, Windows POS is so much better than Linux. It's just the facts.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
