We've run into somewhat of a sticky situation. There are a couple parts to it, so I'll try to lay it out as best I can.

We got a question that was very, very broad:

Could ALT + F4 to close a game corrupt Save data?

As it was, it was well on it's way to being closed. Thankfully, it's been narrowed to reference one specific game. Unfortunately, before that happened, it got an answer. That answer makes some extremely sweeping statements, and is rather problematic at this point in time. Generally, I adhere to the good faith effort rule; a user asked a question, and someone tried to answer it, so it should stick around. Unfortunately, in this case, that answer is dangerously wrong. It's also been accepted, and is very highly upvoted, which adds to the problem.

The answer states that it's absolutely impossible to corrupt a save game by closing a game through Alt+F4. I...find this rather hard to believe. The logic behind the statement makes sense, and from an idealistic standpoint, I even agree with how it should work. Unfortunately, we don't live in an ideal world, and I don't see how it can be true. There are hundreds (if not thousands) of PC games available, from pre-DOS games, to include Linux and other OS's, all the way up to current-gen. Even if we narrow the statement to only include Windows games where Alt+F4 closes the window, I still can't trust the statement. Using the same programmer expertise the answerer did, I come to the opposite conclusion; I cannot believe, with all the amount of code of varying skills, libraries, and tools, that every single game on Windows will never corrupt a save when closing via Alt+F4.

The only source we have for this statement seems to be, "Trust me, I'm a programmer." When pressed for proof, the user refuses to provide any. My very first comment on the answer sums it up fairly well, I feel:

I'm also a programmer, and am very familiar with how simple it is to screw things up. And the vast number of libraries you can code in that handle everything their own way. There is absolutely no guarantee that this is how any game at all handles saving.

So what do we do? Technically, the answer no longer answers the question, post-edit. In an attempt to preserve the answer, I requested the user narrow it to the specific game mentioned. That doesn't seem to be happening.

I personally want the answer deleted, due to the potential harm it can cause by encouraging such a bad practice. But I'm curious to hear your thoughts on this, both for this specific case, and generally on how to handle this in the future. Mods generally aren't supposed to review answers based on correctness, but I'm not exactly sure what other recourse we have here. Allowing the long tail to work on this answer seems to be moving it in the other direction.

  • 4
    I think that regardless of someone's thoughts on the answer itself, it needs to be edited for the new narrow scope. That, IMO, is the bare minimum that has to occur here. If that doesn't happen, I worry we're setting a dangerous precedent. – two bugs Jan 17 '16 at 19:38
  • @twobugs Is it not allowed to talk about the general case? Why not revert the question? – HalfKiloByte Jan 17 '16 at 22:30
  • @HalfKiloByte That would make it Too Broad. And completely close-worthy. – Frank Jan 18 '16 at 0:20
  • 1
    IMO, it's fine to make broader comments about the general case, but you should make sure you answer the question itself. As it stands, the answer in particular doesn't do that. – two bugs Jan 18 '16 at 1:53
  • 2
    found an example where multiple people have issues using alt f4 to close the game: /thread – Aequitas Jan 18 '16 at 2:07
  • 4
    @Aequitas: Ok, makes me happy for you?! In what way its related to this? No one said that this can't happen. Even Manson is actually mentioning that any software could supress or even overwrite the behaving. – Zaibis Jan 18 '16 at 13:03
  • 3
    @Aequitas Did you read the thread? You found a game in which multiple people report multiple problems with save corruption whether or not Alt-F4 is being used. Therefore, it's dubious at best to claim that Alt-F4 causes save corruption on this game; the game's save mechanism is apparently doing a good enough job of corrupting things all on its own. – Mason Wheeler Jan 18 '16 at 19:16
  • 1
    @MasonWheeler The point is irrelevant. You know this. If it breaks in other ways can point to a greater problem, but that doesn't automatically absolve your specific use case from the issue. Alt+F4 can break it. It shows that, and there's doubt about it. Should it? No, probably not. Shows a pretty shoddy level of QA from Namco Bandai, really. But it's a pretty decent proof that Alt+F4 can, and will, break savegames sometimes. – Frank Jan 18 '16 at 22:43
  • 5
    The problem with the top answer, and I too am a programmer, is that it assumes programmers use good programming practices. For the most part the answer is true for all games and programs that are well written, however if fails to mention, unless I'm blind, that if the programmer passes the message to the default handler blindly then the result could be bad. Also some programs don't necessarily handle messages in the order triggered. – ydobonebi Jan 19 '16 at 19:49
  • 1
    I would like to think all programmers of top games at least are experienced enough to implement safe handling, but sadly I have had game corruption as a result of foolishly using Alt-F4 though it could have been coincidence. The game was terraria and I lost not only my active player and world but all worlds and players became corrupted (the data files got changed to all zeros somehow) – ydobonebi Jan 19 '16 at 19:55
  • I know this is about a specific question but what about out of date answers? i recently answered for a bounty because all the other answers are OoD, atleast 3 other answers have a higher vote and 1 of them is accepted, the answers are not harmfull, but they are no longer correct, an issue that happens a lot on MMORPG games because those types of games continuously change their game mechanics. Downvoting is an option ofcourse, but some answers have been upvoted so many times, it would take a lot of community effort to downvote it again, and people up vote more then they down vote – Vahx Jan 19 '16 at 20:38

Setting aside for a second how accurate or otherwise this particular answer is, in a more general case:

  • If an answer is sorta-right, but only in a certain set of circumstances, and doesn't make that quite clear, edit it to make it clear, and maybe prompt the OP to expand their answer if they can. An answer that contains a lot of great research but doesn't quite answer the exact question asked can be improved, and we all have that capability.

  • If it addresses doesn't directly address the specific circumstances of the question, but does address a broader case, edit it to make this clear. All it takes is a line at the top of the answer: "I'm not sure about [specifics], but in the general case:".

  • If an answer is broadly on the right lines, but makes some generalisations that are wrong, edit it to fix them.

  • I hope you get the picture by now.

If you're editing an answer and it's getting into a rollback war, that's another problem entirely that we can handle very easily.

In the case of this particular answer, there's a lot of good information and research here, and in my opinion it is perfectly reasonable to expect a big-budget, AAA game to implement safe handling of ALT-F4, and an answer explaining that the exact behaviour can never be known except by the developers (but it's reasonable to expect that this is how it works) is perfectly valid.

This particular answer seems to me to be 90% on the money, and needs a little coaxing to make it perfect.

  • 14
    We are not mathematicians, we do not require absolute proof. As I said, I think it reasonable to assume that an AAA game implements Alt-F4 correctly if there is no evidence that it does not. – fredley Jan 17 '16 at 22:19
  • 12
    @Frank Nothing "cuts it" for you, does it? My central point is not "trust me, I'm a programmer"; it's "if this were really a problem, it would really be a problem, but there's no evidence that such a problem exists. And here's the reason why it's not, and as a programmer I'm doing my best to explain the theoretical details behind why it isn't a problem." That should really be all the "proof" anyone needs, but for some bizarre reason, you just can't let it go. – Mason Wheeler Jan 18 '16 at 2:17
  • 6
    @MasonWheeler You've hit the nail on the head with your statement: Your answer is theoretical. We're about definitive facts. You're completely guessing, and that's anethema to SE, which you know. As is, a thread has been found that serves as a perfect example as to why we don't want guesses. Tales of Zestiria doesn't like Alt+F4 – Frank Jan 18 '16 at 4:22
  • 4
    @fredley Please don't encourage him; not when his attitude makes it abundantly clear that any edit he actually makes will end up being an act of vandalism. :( – Mason Wheeler Jan 18 '16 at 9:15
  • 7
    @MasonWheeler Vandalism is something we can deal with very easily, however I'm keen to see Frank take some positive action. – fredley Jan 18 '16 at 9:19
  • 7
    @MasonWheeler You keep moving the goalposts. Someone shows you saves actually corrupt, and then you go, "Of course that'll break if this was happening!" which runs counter to your entire point. Saving can, and will break. Logic plays no part in it; things screw up because we are imperfect. Your inability to admit that damages your own credibility and contributions. I don't think I'm able to edit that answer in any manner that would both be correct and still keep intent. – Frank Jan 18 '16 at 13:08
  • 6
    After reading all of these threads, the original answer and chats it is pretty apparent that Frank's behavior is harassment. I think until he can provide a valid counter example he shouldn't press this issue further. – hellyale Jan 18 '16 at 13:15
  • 6
    @Frank Trying to improve an answer is a noble goal. Taking an ad-hominem approach is not the right way to do it though. – fredley Jan 18 '16 at 13:24
  • 5
    @Frank hit alt + f4 mid save on the game in question, reloaded game , no issue . Repeated expirement 5 times. No issues. Can you provide a game that has issues on an alt + f4 close that doesnt already have save corrupt issues from other causes? – hellyale Jan 18 '16 at 13:27
  • 4
    @MasonWheeler In a week everyone will have completely forgotten about this, and then I'll go through and clean it up. For now though it's best to not mess with the discussion. – fredley Jan 18 '16 at 17:23
  • 4
    @MasonWheeler I believe I understand your point quite clearly. You're saying that Alt+F4 never has, and never will, corrupt a save game. That's quite a sweeping statement to be making, and the whole reason for this post. Logically, yes, it makes sense. I totally agree with that, even. From an experience standpoint, however, I know firsthand how easy it is for interlocking systems to conflict in strange ways and lead to very strange outcomes. I find it extremely hard to believe that that would never have happened, in the thousands of games created in the last few decades. – Frank Jan 18 '16 at 22:50
  • 6
    @Mason Please remember your post is a wiki. SnakeDoc's addition does not, as far as I can tell, conflict with your intent, it clarifies it. As a mod, you should know better than to engage in rollback wars. – fredley Jan 19 '16 at 21:05
  • 5
    @MasonWheeler You keep acting like a the only possible way for a save to become corrupt with ALT-F4 is if the programmer had some sinister urge to corrupt save files. "if the developer deliberately sabotages the program to cause it to do so" is such an absurd statement to make, especially from another developer. I don't believe anyone intends to corrupt data - but it happens all the time. Its very easy to do wrong - there is no standard windows API saveGameAndExitNormal() method that handles this automagically. Its up to the developer, who will have varying skill and knowledge. Its possible. – SnakeDoc Jan 19 '16 at 21:24
  • 7
    "Any developer who knows enough (or team that collectively knows enough) to produce a complex game and bring it to market will know that." You are making huge assumptions. This is exactly why your answer is so disputed. In the best case, you may be right, but we all know the best case is rare. Have you seen the quality of code produced by big shops? I've never seen a AAA title's source, but big pro software vendor code will frighten you senseless sometimes. The point is, it's more possible than it's not, and therefore you answer as-was is misleading. The disclaimer stands. – SnakeDoc Jan 19 '16 at 21:53
  • 4
    There are plenty of programs, including games, that are "brought to market" by a development "team" that is learning on the fly. This often includes plenty of instances where they reinvent the wheel in various places, or hijack certain parts of the process that the framework they're using, or even the OS, would normally control automatically. It happens. – Ellesedil Jan 19 '16 at 22:05

I've seen two complaints about the answer. One is that the answer is wrong. The other is that the answer is harmful. Let's look at those:


We don't even need to look at the specific question/answer to evaluate this one. It's simple. We don't delete wrong answers. We've never been in the business of deleting wrong answers. People can self-delete their wrong answers. Everyone else can downvote them. Note that in the deleted answers section of the site help, there's no "it was wrong" delete reason.

If someone really feels a downvote isn't enough (because a wrong answer is highly upvoted), they can leave a comment explaining why the answer is wrong. Anyone reading the answer will also read the comment and can at that point decide whether they put more stock in the upvotes or the comment explaining why all those upvotes are wrong.


Answers that are actively causing a hidden harm should, in my opinion, be deleted. In most of the cases where it happens, I'm betting it would be someone actively trolling. (i.e., "Q: How do I save my game?" "A: Go to the command prompt in Windows and type format C: /u").

Notice that I said "hidden harm" though. Looking at the specific question that spawned this, there's no hidden harm. The question is specifically about whether or not alt-F4 is harmful to a save. If the answer of "no" is wrong, it's not causing you some kind of surprise harm. It just means the answer about whether or not it was harmful was wrong, and you can now downvote it and leave a comment about how it was wrong.


It doesn't matter if an answer is wrong. That's what downvotes are for. If the answer is not harmful beyond the extent of possibly being the wrong answer to the very harm the question is about, no action is needed.

  • 1
    "Leave a comment" is no good. The mods have specifically prohibited comments on this answer because it's generating too much debate. – Kevin Jan 18 '16 at 16:59
  • 2
    @Kevin In the general case, posts don't blow up into drama hurricanes, so leaving a comment is a good thing to do. – fredley Jan 18 '16 at 17:02
  • 1
    @fredley: I thought we were talking about this specific case... What are we supposed to do when commenting is disallowed, anyway? Your answer does not address that. – Kevin Jan 18 '16 at 17:03
  • @Kevin As it says on the post in question, either take it to chat which is much better set up for extended discussion, or take it here to meta. – fredley Jan 18 '16 at 17:04
  • 2
    That's precisely what Frank did, and now he's getting downvoted and ignored. – Kevin Jan 18 '16 at 17:05
  • 6
    @Kevin There's a big difference between leaving a comment about the problems you see with an answer and having long an argument about it. Usually that difference comes down to tone (which can make the OP defensive) and whether or not you continue to respond to arguments about your comment. The comment isn't to convince the OP that they're wrong... it's to give other users some reasons why the OP might be wrong. Besides, given that the post in question now links to this meta, it's still effectively raising fbueckert's objection. It unfortunately just requires some clicking now. – Sterno Jan 18 '16 at 17:17
  • I should be clear I'm not saying he did anything wrong. The comments were deleted long before I ever read them, so I have no idea what their content was. But it is certainly not unheard of for long comment chains to get deleted and discussion moved to meta. I don't think that happening here is an argument against leaving a comment when you think an answer is wrong. – Sterno Jan 18 '16 at 17:22
  • @Sterno It's worth pointing out that in many cases, including this one, the answer can be edited to remove the parts of it that are wrong. – fredley Jan 18 '16 at 22:51
  • 1
    For the most part, I agree with this, except for the harmfulness of it. I do think there is a fairly large potential for harm, if this is taken as gospel across all of gaming in general. Specifically, probably not, though, and that's probably how it'll be interpreted for the most part. – Frank Jan 18 '16 at 23:05
  • 3
    I thought it was cool you could save any game from the command prompt so I tried it and now I can't get my computer to open, what gives? – Aequitas Jan 19 '16 at 0:59
  • 1
    @fredley I think the fact that this answer "blew up into a drama hurricane" is indicative of the issue. As written, the answer is likely to leave false impressions with readers that it's impossible to corrupt save games by ALT+F4'ing out... this an impossible claim to make and impossible to prove. It's far better to air on the side of caution and advise users to instead utilize the in-game intended exit routine. Without that disclaimer, the answer should be considered dangerous because it will cause someone to corrupt an important save. – SnakeDoc Jan 19 '16 at 17:35
  • 1
    @SnakeDoc This is why this argument seems so ridiculous to me. The guy is asking if it is harmful. Some answers say yes. This one says no. People are up in arms that this "no" answer is going to get him to hurt himself. It's specifically the thing he's asking about. If he gets a yes answer and a no answer, it's not like he isn't fully aware of the possibility that the "no" answer might be wrong. We need to stop treating users like idiots who can't make decisions for themselves. There's no hidden danger here. He's fully aware of the risks if the answer is wrong. – Sterno Jan 19 '16 at 17:55
  • 2
    @Sterno Nobody will be physically injured by a corrupted save, but it's parallel would be as-if we had questions about Snow Camping and the top answer said "wear cotton!" and got a ton of upvotes. It's dangerous in the context of the SE site. The problem many have taken with the answer is Mason's stubbornness to at least even admit there is a remote possibility of corruption. The more it was debated, the more adamant he became that there was no scenario which could prove him wrong. His ego probably got the best of him, but that's an absurd claim to make. The answer has been updated now. – SnakeDoc Jan 19 '16 at 18:00
  • 3
    @Sterno Essentially, without being able to look at the source code for every game ever made, nobody can make any claim that under no circumstances will ALT+F4 lead to save corruption. It's more wise to say that it's unlikely to lead to save corruption in a modern, well designed game, but that it still has the possibility and therefore it's still best to use the in-game exit routine because that's how the developers would have anticipated a user quitting. Mason's claims are well reasoned, and are how a game ought to be designed, but he cannot make the claim about every game. – SnakeDoc Jan 19 '16 at 18:04
  • 2
    @Sterno It's not a "yes or no" question, it's a "maybe or sometimes", however the answer has taken the only "no" route, which is wrong. It is closer to my original analogy, if we had a question about wearing cotton in the snow, and someone started off by insisting there's absolutely zero risk ever, they would not only be wrong, but they would be spreading potentially dangerous information. That's the case we have here. – SnakeDoc Jan 19 '16 at 20:34

The answer states that it's absolutely impossible to corrupt a save game by closing a game through Alt+F4.

Instead of absolutes, the answer comes from a practical viewpoint. If the software is really liable to get Alt+F4 wrong after Win95 had it working correctly in MSPaint, any other part of it is of questionable quality too. If the software is really so bad, then it's highly likely that some other feature would cause corruption even without trying Alt+F4.

The answer should not be deleted because "it's dangerous". At most, it should include mentions for both the safe and the casual approach, and let the reader decide.

  • 1
    Pragmatic or not, it is wrong. In an ideal world, yes, that is how it would work. We don't live in the ideal world, though. – Frank Jan 17 '16 at 21:28
  • 2
    that's not what the answer states though, they say it's impossible unless the developer is an idiot – Aequitas Jan 17 '16 at 21:43
  • 9
    @Aequitas Speaking as a developer, all developers really are idiots from time to time. – fredley Jan 17 '16 at 21:51
  • 8
    @Frank Please consider giving the readers options with well reasoned explanations and letting them decide, instead of deciding for them. – HalfKiloByte Jan 17 '16 at 22:28
  • 3
    That's exactly what Meta posts are for. – Frank Jan 18 '16 at 0:20

Very highly upvoted, very wrong answers get very deleted. We have one precedent for this and it was a rather uncontroversial call.

It's simple to see why: we're spreading misinformation and that's not cool. This should never happen, but occasionally it does, and when it does happen we don't really have anything else we can do about it.

This isn't the case, however: the answer is not wrong to the extent that Gnome's post was. There is no reason to resort to such exceptional means.

The community rightly felt that the advice was dangerous to rely upon, so we wanted to stick a warning somewhere. I moved the warning and put it to the question so that, no matter what happens, the warning will appear in the page before all answers. In an ideal world I would've put it in a custom post notice, but I can't do that, so it's in the question body itself. Several answers to the question do, after all, agree that Alt-F4 is safe to use, and it would be silly to stick a warning in all of them, let alone just one of them. I also added a few particular cases that I felt were not addressed by any of the existing answers.

I hope this resolution addresses everybody's concerns.

  • 3
    You just had to make it about porkchops, didn't you. – fredley Jan 20 '16 at 10:21
  • 1
    You've also made the question too broad again. – Frank Jan 20 '16 at 13:01
  • @frank the question is as broad as the answers are. – badp Jan 20 '16 at 13:39
  • @Frank / badp - I've been generally lurking throughout this debacle, but the question & the answers smacks of developer intent. In the general case the answer is "Generally not, but it depends on the devs of each game", and in the specific 'Saint's Row' case "It depends on how the devs of Saint's Row coded it". I'd honestly consider closing on those grounds tbh. (Although I find 'Too Broad' a bit of a stretch). – Robotnik Mod Jan 20 '16 at 13:50
  • 4
    @robotnik "it depends on how it was coded" is not a question of intent. It's a question of exactly how a thing works. It's what we do here, to some extent or another. Developer intent is about just that - intent - as in "why did the Saints Row devs make it so that hitting alt-F4 can corrupt my save? Are they incompetent, or simply malicious ne'erdowells?" – LessPop_MoreFizz Jan 20 '16 at 16:42
  • That disclaimer mostly solves the problem, good work. – DCShannon Jan 20 '16 at 23:27

In the first point: Ofcourse this shouldn't be delted just because its partial wrong.

But about its discussion: It was never absolutely stated that this can't happen (As you actually claim.).

He just said that this is not to be expected in case there is no inappropiate coding done.

And your counter claim is, that we have to expect inappropiate or hasty behaving since this isn't a perfect world.

But this moral would mean conclusively, that we can't rely on, what software is supposed to do. So we should deny information technics.

What would mean that you both are wrong and paradoxical also no statemant about software can ever be true.

What mason actually did, is he just took refference to documentation of features and coding conventions.

And you say he is wrong about it since there are programmers who won't use best practice coding conventions, or who accidently will code bugs.

But thats exceptional and not a rule. Also if you read a feature/API documentation you can asume to be that way. Otherwise you shouldn't even be able to install windows since you can't be sure it is doing what it is supposed to do.

Mason asked for a counterexample, what you couldn't provide (and even you could, it wouldn't proof anything since implementing bugs in software is anything but not a rule).


Your whole arguing is paradoxical relying on software can't do what it is supposed to do since humans make misstakes.(point granted but can't be converted into a rule!)

Masons answer is relying on primary sources so assuming this is the way it works is more valuable then assuming it isn't because there could be bugs for the definition of what he writes beeing true.

Apart from that:

His Answer beeing wrong or right doesn't matter. It is a try to answer the question which even is addressing the OP in any aspect. So deleting it and even escalating it to meta is inept! It is ON-Topic and therefor its value of beeing usefull has to be determined by votes.

  • ...Huh? Your point is hard to understand. That said, the onus is on Mason to prove his point. You don't get to make sweeping statements like that, and then decline to prove it when challenged. – Frank Jan 18 '16 at 13:10
  • 1
    @Frank: you should reread his post. He is just refering to the documented point how things are expected to be. he even notes that this can be altered by bad purposes or accidently. So there IS nothing to proof... even dealing wiht examples wouldn't be an proof since they are out of the features itnention scope anyway. In other words: In case a verified primary source (what microsofts documentation about alt + F4 in this case would be). Then there is no way of being incorrect by just refering to it. Since any coutnercase just is in itself wrong by not beeing conform to primary source(docs) – Zaibis Jan 18 '16 at 13:32
  • But...that's sort of the point; he's dealing in theoreticals. Logically, yeah, I totally agree that's exactly how it should work. I really do. Unfortunately, I've seen what we as coders can do, even when we're focused on doing it the best way we know how. We also work in the real world, which almost always never matches the exact ideal. I am extremely doubtful that every single use-case can hit the exact same behaviour every single time. That's why we have QA, unit tests, code review, and whole bevy of other eyes on our code. We make mistakes. – Frank Jan 18 '16 at 22:40
  • "He just said that this is not to be expected in case there is no inappropiate coding done." This is called a bug. Bugs happen. More-so when developers are learning their craft on the fly. If a developer is tasked to work on exit logic or saving logic, and they've never written that particular functionality before, they'll try to implement it as best and/or quickly as they can. Researching best practices would be ideal, but doesn't always (I'd say rarely) happen. And that's why Mason is wrong. He can't guarantee that alt-f4 saving will always work because he can't guarantee bug-free code. – Ellesedil Jan 19 '16 at 22:15
  • 1
    @Ellesedil: He just actually ISNT guaranteeing anything... what the heck... – Zaibis Jan 19 '16 at 22:19
  • Technically true. However, his only exception "would require the developers to quite deliberately go out of their way to screw up the saving system." For starters, the implication here is that developers wouldn't intentionally sabotage their own code (a safe assumption, but not a 100% guarantee), which is why Mason's answer reads like "all games ever will never corrupt save data when exiting with alt-f4". – Ellesedil Jan 19 '16 at 22:31
  • Second, this isn't correct either, and it is actually somewhat insulting, particularly towards developers who simply do not know better. At some point, developers need to learn how to implement exit code and save code, and we usually learn by doing. Programs are complex and making a change in one spot could affect code in other places, including code that isn't directly related to the changes being made. A change being made in the exit code could very well effect how data is saved. – Ellesedil Jan 19 '16 at 22:32
  • It would depend on the program's implementation. That's simply a bug, or a failure in design. But it's not some deliberate intent to sabotage anything. – Ellesedil Jan 19 '16 at 22:32
  • 1
    @Ellesedil: are u all just trolling around? man and if the universe crashes into another universe in our multiverse and all collapses within the fractal of a nano second and exactly in that moment OP tryed to save his game his save m8 be by some definition also corrupted..... But do we have to respect this on an answer we give? No. because in that case things are fucked up anyway. And same is here.... Do we have to respect a programmer who intentionally sabotages code, while we explain a feature? NO!!! And why? BECAUSE IN THAT CASE THINGS ARE FUCKED UP ANYWAY – Zaibis Jan 20 '16 at 4:47
  • You clearly didn't read or understand my last set of comments. I'm disagreeing with the insinuation that the only way a save could be corrupted is by a developer intentionally sabotaging code. People make mistakes all the time, particularly while learning. It doesn't take any nefarious intent to write bad save or exit code. But that is what Mason explicitly says, while strongly implying that most developers wouldn't do that. It is insulting to developers who haven't refined their skills, and it is wrong. – Ellesedil Jan 20 '16 at 4:52
  • @Ellesedil. I understand you point but I can't read in his answer what you are reading there.... He is just not taking this in to account, same as he isn't taking into account what affects the universe collapsing would have onto it. (thats what I tryed to refer to with my last comment). So just because he isn't stating it doesn't mean actually he is expressing this can't happen... He just doesn't talks about it. And I reread his post... Tell me where he explicitly states that developers wouldn't do that?! I just can't find it, even mroe I jsut find the opposite. – Zaibis Jan 20 '16 at 5:06
  • He literally says: "actually doing so would require the developers to quite deliberately go out of their way to screw up the saving system". Since this is the only condition he places on save files not being corrupted, he is wrong. He ignores developers that simply write bad code because they lack the knowledge, skills or time to do otherwise. And that's the whole point of these two days worth of posts. He paints writing code like this as intentional and ignores the fact that not every developer is highly skilled/experienced and that humans make mistakes. I'm not sure what else to tell you. – Ellesedil Jan 20 '16 at 5:26
  • In any case, I've wasted enough time on this topic. He's a developer, and other developers such as myself have seen plenty of bad code that do things unnecessarily to know that what he's saying cannot automatically be applied to every single program in existence, game or otherwise. I'm content with the disclaimer on the answer, because it's otherwise a decent answer. If you can't understand this position, then there's not really more I can do to help you. – Ellesedil Jan 20 '16 at 5:30
  • @Ellesedil: Then thanks for your time. Its mayber caused yb my english skills, but the quoted sentence you refer to doesn't imply for me what you say it does. but anyway your right we should stop wasting time on this. – Zaibis Jan 20 '16 at 8:24

Agreed (That it should be removed/hidden) - or a large warning edited onto the answer warning anyone who may believe it, caveat emptor due to the unreliable nature of the 'facts'.

Infact it wouldn't be impossible to write a demo program that proved exactly how easy it is to corrupt data by handling a signal badly (which is what the answer in question claims isn't possible).


Nothing should "be done". This answer was accepted and highly upvoted because it's correct, and mods aren't supposed to review specific individual answers for correctness. You've been harrassing me for 3 days over this. Knock it off already. If my answer were as problematic as you say, there would be examples of it all over, but no one has presented a single counterexample, because the existence of one would violate major, well-understood (except apparently by a handful of trolls on Gaming.SE) principles of programming. Just let it go.

  • 5
    Thank you for demonstrating exactly the same behaviour as you did on the answer. Your answer is wrong. You refuse to prove it. Therefore, we have to do something about it. – Frank Jan 17 '16 at 20:05
  • 4
    @Frank What part of "don't ask someone to prove a negative" do you not understand? My answer is obviously correct to anyone who understands concurrency. Provide a counterexample or you're doing nothing but straight-up trolling. – Mason Wheeler Jan 17 '16 at 20:07
  • 1
    If this cannot happen, you can show us proof. Any proof. It doesn't have to be of a game not being able to corrupt a save. It could be something like making a small proof of concept where you prove everything gets funnelled through the WM_CLOSE command. And has for twenty+ years. The point is, if you're going to make a statement, the onus is on you to prove it. Doesn't have to be concrete proof, but at least an effort should be made. – Frank Jan 17 '16 at 20:13
  • 15
    @MasonWheeler I am, like Frank, a programmer. In fact, I'm a Windows programmer. More than that, I'm a Windows GUI programmer. I know what can happen when processing an event like WM_CLOSE. The fact is I've seen cases where the program hits an exception and drops out of the event routine half way through without any indication to the user! If you hit such an exception half way through a save (unlikely, but possible), then you've just corrupted that save. – MBraedley Jan 17 '16 at 21:17
  • 15
    Your answer was not necessarily highly upvoted and accepted because it is right, but because it sounds about right. The thing about Arqade is that most users are not programmers and a lengthy answer going into technical details in an unknown field attracts upvotes just because "this guy seems to know his stuff!" (I know my own upvote came from that direction). However, Frank and MBraedley are raising valid points, and considering they both bring the exact same credentials as you do, and are adamant that your answer is wrong, there is indeed something to be corrected here. – MrLemon Jan 17 '16 at 21:33
  • 2
    I'm curious, in your answer you kind of claimed that the save thread must be the only one modifying the game data, or else race condition might happen (if the other thread happens to be modifying the game data), and this would result in frequent save corrupt. But couldn't the message handling thread be the only other thread running together with the save thread? Then if the programmer had programmed WM_CLOSE to immediately close the window, then it will be immediately closed, and save might be corrupted (which won't happen if you don't close while saving, so no frequent corrupted save) – justhalf Jan 18 '16 at 9:27
  • 2
    Mason, Frank can be abrasive and absolutist, and often refuses to admit he's wrong. He's usually right, though, and he is this time. You need to recognize you're acting the same way. Repeating your own advice: let it go. – DCShannon Jan 20 '16 at 23:32

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .