War Games/Headscratchers


 * Hasn't the ending pretty much left the US defenseless against a real nuclear attack? If it came, the computer would want to play chess...
 * There is no practical military defense against a real nuclear attack as it happens. America's nukes won't defend bupkis in the event of an actual nuclear attack; the defense they provide is in convincing potential attackers that it's not worth it to fire in the first place. And anyhow, now that WOPR isn't hellbent on "winning" the game scenario started by Lightman, removing it from control of the missile system is probably not a problem. So, no, the USA isn't any more defenseless than it was before the movie started.
 * In fact, it's probably more so- since WOPR won't start a nuclear war by accident.
 * There's also that they can, and almost certainly will, go back to using manual control on the missile silos.
 * So why, after going through the Falkan's Maze "security" feature, did it activate the option of starting a thermonuclear war between the U.S. and Russia? If you play as Russia, the screens show Russian missiles attacking the U.S., and presumably, if you play as the U.S., it launches an attack on Russia. Therefore, the only conceivable use for the system would be to let Falkan tell them "I told you so" in the hours before complete and utter destruction of a third of the human race. It has been suggested that they just didn't know to separate the Global Thermonuclear War "game" from the weapons defense system after setting up the supercomputer for missile detection, but it seemed to me that the programmer did it on purpose.
 * The movie explains all of it. First off, missile launch was originally done individually by silos, with launch officers turning keys. In the movie they switched away from this system after a full-scale dress rehearsal showed that too many of the silo launch officers were not psychologically capable of actually turning their keys. (That's the opening scene of the movie, remember.) So they rewired the silo launch controls to the central control panels at NORAD HQ, with two safety measures: 1) the ten-digit launch control key and 2) the missiles could not be launched unless General Beringer ordered DEFCON 1. WOPR used a series of false alarms to trick General Beringer into setting DEFCON 1, at which point it did a brute-force hack of the launch key. Or to put it more simply: nobody actually gave WOPR control of the missiles, it did a social and then a crypto hack on the NORAD HQ systems.
 * But when WOPR started trying to brute-force the launch key, they clearly noticed it and wanted to prevent it. So why didn't anybody order a step-down from DEFCON 1? Wouldn't that have prevented the launch?
 * WOPR was by that point no longer responding to commands. Note that they don't reset DEFCON to 5 until after WOPR has been convinced to relinquish control of NORAD's computer systems.
 * This is actually in the movie as well; General Beringer orders one of the operators to lock out future changes once the situation he's convinced is happening for real gets to what he feels is the point of no return. When they try to stand down, before realizing that WOPR is hacking the launch code, there's even a brief shot of the blinking light that pressing the lock out button turned on to remind you of it, suggesting that this is the final thing WOPR needs to act of its own accord. This is in tune with how the entire problem is caused by removing humans from the loop, it's the logical conclusion; once they hit the point of no return, human input is no longer needed. If this seems like an illogical extreme, remember that this is the eighties, or at least the eighties as viewed through the lens of a populace suffering from, if a small degree of, nuclear panic and Red Scare in their daily lives. The humans who set up the system would, at some point, think "What if NORAD, in a million to one chance, is infiltrated by a Soviet agent? Even if it means a suicide run, they could, possibly, interfere with our nuclear response. Better make a point-of-no-return switch that takes the humans in NORAD out of the loop too." To say nothing of the idea that since the missile commanders had problems going through with it, who's to say that General Beringer wouldn't, and if he fails to order changes locked out, the new doctrine would be to assume that he's not up to it before it comes to the actual launch?
 * As to why it didn't do that until after it was hacked? That's more speculative, but this troper's guess is that Matthew Broderick's character, in his ignorance, made the mistake of issuing WOPR an open-ended command that WOPR interpreted as being given permission to do what it did. Which makes this less of a story about a computer that went crazy, and more the ultimate cautionary tale about Garbage In, Garbage Out.
 * When Broderick brings the girl over to show off his mad computer skillz, he tells her not to touch the keyboard. She says she won't, and the leaves streaks of fingerprints down the screen. It's not a plot hole, but it really, really bugs me.
 * Well, maybe Broderick is the kind of guy who doesn't mind people touching the screen (as that can just be cleaned), but doesn't want people touching the keyboard and typing in something stuPENIS.
 * The computer was auto-dialling at the time, he just didn't want her to terminate the program by accident.
 * It's been a while, but why does NORAD's supercomputer have Global Thermonuclear War installed on it?
 * Mc Kitrick probably didn't know about it. I haven't seen it in awhile either, but I seem to recall that part of the problem was Falken, out of grief for his dead son, putting a lot of crap into that thing that wasn't supposed to be there. Nobody believes David at first when he starts, seemingly out of his gourd, ranting about how Joshua is playing games, so they probably didn't even realize the WOPR was apparently sentient.
 * Joshua was a computer designed to make the "perfect" moves in case of a global thermonuclear war, faster than human operators could think or react. The description of Joshua's system is basically a neural net - you train it up on simple tactics using simple games, eventually culminating in a simulation that matches what the computer would be doing in real life. Joshua just got confused near the end about what was real and what was a game.
 * WOPR was meant to run attack/defence simulations. A real-world simulation is called a "wargame". So the simulations were put under 'games'.
 * Who's idea was it to pull that alarm prank on the tour group? Someone's getting fired...
 * Probably the same person who thought civilian tours through high-security military bases would be a good idea.
 * That bugged this troper too. Those boards are going to have all sorts of classified information on them, such as silo locations, not to mention the VIP bunkers. Why are they letting civilians see them?
 * This troper would just like to point out that nuclear missile silos aren't generally hidden in American doctrine - they were in the USSR, but not in the US, where they were usually sited in the middle of absolutely nowhere for a few tactical reasons
 * But they'll see the big board!
 * It's really not that strange. This Troper's father worked on another Air Force Base in Colorado Springs, but he had no trouble getting a tour arranged for me and my elementary-school-aged friends. And yes, we saw the big board.
 * How was Matthew Broderick's character able to hack into a Do D mainframe using just a Commodore 64?
 * Not a Commodore 64 - this was released in about 1985. He was running an IMSAI Intel 8080-based S100 box; released in the late '70s as a copy of the Altair.
 * There was a little more to it than that, obviously; he had a ton of other hardware attached to it (IIRC, Word of God holds that he got most of it by Dumpster-diving). Plus, he's pretty damn tech-savvy.
 * Not to mention that the mainframe's original programmer left a great big whacking backdoor hole in the security; an unsanctioned, undocumented hole.
 * Building on that last one: why did the Department of Defense have one programmer write both the AI and the security framework for the computer on which they were to run? These are very different tasks, and it's unlikely that any programmer who is as good at AI as Falken is supposed to be would be proficient enough to write military-grade security protocols.
 * Remember, 80s. This is before the invention of software firewalls. This is when you dialed-in directly to a mainframe. The 'security framework' is Joshua's own authorized user list, which Dr. Falken obviously has access to, since he has a root account. Joshua's primary 'security framework' is that you are not supposed to be able to even get to a terminal connected to the machine, as its NORAD-internal only. Again, the ability to dial-in from an outside trunk line was a grave switching error.
 * Security framework in this case includes, but is not limited to login/password. The system would have to be modified to allow certain users to log in without a password, which would never happen on a timesharing system, if for no other reason than preventing people from tying up the phone playing Spacewar! after they had used up their time budget. Allowing this sort of a backdoor would require rewriting large portions of the operating system.
 * Aaaaaand, who wrote the kernel for Joshua OS again? Professor Falken. Shazam.