Bill Gates Acknowledges Ctrl+Alt+Del Was a Mistake

theodp writes "If he'd had his druthers, Bill Gates told a Harvard audience, Ctrl+Alt+Del would never have seen the light of day. However, an IBM keyboard designer didn't want to give Microsoft a single button to start things up, and thus the iconic three-finger-salute was born."

Comments

Re:Redundant keys

[HaZardman27]

I find that at least I use both of the shift keys, unlike the right Ctrl, Alt, and Windows keys.

XT was a mistake

[ruir]

if we start thinking in what was wrong building a computer with 25 years technology then we would lost count of the mistakes. If it weren't for the backing of IBM, it would been an instant market failure.

Re:XT was a mistake

[Stormwatch]

Not the XT, but some food for thought:

Back when IBM was planning their original personal computer (note that "PC" used to mean any personal computer, but the IBM-PC's success hijacked the name), one design they considered was based on the 801, an early RISC processor about 45 times more powerful than the 8088 they ended up using, and their own modern operating system instead of Microsoft's junk... yeah, that's no typo, FOURTY FIVE TIMES more powerful: the 8088 @ 5 MHz did a paltry 0.33 MIPS, while the 801 did 15 MIPS.

Think about it, the PC started with the wrong foot. A broken, twisted, crippled, atrophied foot. How much technological progress did we lose because we had to waste time overcoming the limitations of this super shitty beginning?

Re:XT was a mistake

[DarkOx]

How much technological progress did we lose because we had to waste time overcoming the limitations of this super shitty beginning?

Hard to say. Because: Its also true the XT they shipped was built of mostly COTS parts. There is no reason to think a much bigger project the 801 based machine with some operating system that would have to either be written or ported from 390 or elsewhere would have ever made it out the door before getting canceled.

I don't even know your more powerful machine would have been successful; honestly IBMs experience with Operating Systems and interfaces at the time would not have translated well to home users and SMBs. MS-DOS was something motivated people could figure out on their own with manual.

IBM would have had to build something that did not use JCL would would have been radical for them at the time. Because there aint no way people would have picked that up.

So its very likely we would have never had a more or less 'open' PC platform at all, it would have all been the very proprietary 8-bit and early 16-bit stuff in the space from Ti, C=, and others.

We might still be well behind where we are today.

Re:So why is it used in Windows?

[jawtheshark]

Because Ctrl-Alt-Delete is non-interruptible. This way one could be sure it was truly the login screen and not something impersonating the login screen. At least, that's how I remember it. Could be urban legend.

3-fingers are just fine

[Alomex]

In my book that is the one thing they got right. It is a cumbersome combination as it should be since you do not want to reboot your computer by accident.

It still irks me how easy is to accidentally shutdown your computer in windows when all you are trying to do is putting it to sleep through the menu.

In programming languages this is called "syntactic salt" and it is used to implement powerful primitives that should not be used lightly.

Re:BSOD

[SJHillman]

Would your rather your PC just turned off without any error message whatsoever? The BSOD is a useful tool... the mistake that causes it lies elsewhere.

Re:Makes sense

[moronoxyd]

I had a keyboard once with a dedicated start/shutdown key.

After shutting down my system a few times accidentally I threw that keyboard away.

Re: So why is it used in Windows?

[bws111]

Huh? NMI is a hardware thing, Ctrl+Alt+Delete is entirely a software thing. The only thing that ever had to do with NMI (related to this) is that on the PCjr the KEYBOARD used the NMI to signal a keystroke. This had an advantage that even if your PC somehow wound up in a interrupts-disabled state the keyboard interrupts would still be processed, and thus Ctrl+Alt+Delete would still work (the BIOS recognized the sequence and branched to the 'reset' code). On the other hand, it was a mistake because typing could interfere with timing critical things (like async comms). As far as I know, the PCjr was the only machine to ever use NMI for the keyboard.

Maybe what you are thinking is that there was no way (in Windows) to 'hook' the keyboard in a manner that could intercept Ctrl+Alt+Delete. That would prevent things from taking over the logon screen.

Re:Revisionist history

[Voyager529]

Using Ctrl-Alt-Del to trigger login gives you two kinds of security:

1. Software cannot simulate a Ctrl-Alt-Del in order to play games with the login screen.

2. By first pressing Ctrl-Alt-Del, the user logging on can be quite sure that they are giving their login credentials to a genuine Windows (or whatever OS) login screen, and not some malware that merely resembles the login screen.

Perhaps I'm simply misinformed and the software does something different somewhere...but I've 'simulated' Ctrl+Alt+Del from Remote Desktop, LogMeIn, TeamViewer, and VNC...I still don't follow how this still holds true.

Power key was more sensible on Macs

[Valdrax]

I had a keyboard once with a dedicated start/shutdown key.

After shutting down my system a few times accidentally I threw that keyboard away.

Seems like a bad design. Macs had a power key for ages on their keyboard, but it pulled up a shutdown prompt instead of killing the whole machine instantly. (You could hold it down for 3 seconds for a force power down, IIRC.) It was also far above the keys and hard to accidentally hit on the machines I remember. This is the one I had on my Performa 5200, and this was the one my old iMac had. (You can see the power key on the latter above the divide between the letter and numeric keypad sections.)

What was the keyboard you used like, and what OS was it for?

Re:Revisionist history

[DickBreath]

Those programs and protocols have provisions to simulate a Ctrl-Alt-Del. A VM will actually simulate the non maskable interrupt. If you are remoted in to a GUI that is not on the physical console, then there is no keyboard to generate an NMI and the GUI you are connected to simply reacts to the simulated NMI (aka Ctrl-Alt-Del).

But try 'simulating' a Ctrl-Alt-Del for the GUI session attached to the physical display/keyboard. I'll wait.