The Great Duck posted the following question after nearly a week in the Sandbox:

It was closed very quckly as "unclear what you're asking." Considering the number of upticks, it seems most of those votes were following a comment by RonJohn which reads:

However, there are still two serious problems. #1 easily fixed: The title ("constructing a simulated universe to start a war?") doesn't match the body ("a single world-war simulation"). #2 Maybe because I've been a computer programmer for so long, and lived through the Cold War, but the question "what quality-analysis tests would justify using a single world-war simulation?" has such a blindingly obvious answer (your nation-state wants to determine if it would win a WW) that he must be asking something else. Thus, still Unclear What You're Asking.

I corrected the title mismatch (an oversight), but people are following what appears to be an error on RonJohn's part.

but the question "what quality-analysis tests would justify using a single world-war simulation?" has such a blindingly obvious answer (your nation-state wants to determine if it would win a WW) that he must be asking something else.

"Your nation-state wants to determine if it would win a WW" is NOT a "blindingly obvious answer" to the question "what quality-analysis tests would justify using a single world-war simulation?"

It would be a blindingly obvious answer to "why would you write this software in the first place?"

At the time of this posting, RonJohn's comment had 4 upticks, suggesting at least four other people don't understand what RonJohn said and why it's in error.

People are voting to close this question as unclear because they don't have enough experience in the field to comprehend the question. It's analogous to a teenager casting a vote to close a question about celestial mechanics because the teen doesn't understand celestial mechanics. That doesn't reflect a weakness in the question, but a weakness in the user voting to close the question.

There is one fault that's mine

The original question in the sandbox was very lengthy and included a lot of information not germaine to the question, but in my robust house cleaning I threw out one piece of information that likely helps.

  • Great Duck's story is being told from the perspective of the simulated AI. In other words, The AI's world is influenced by the reason the developers chose to use a world-spanning war to test the software.

However, that doesn't actually change the question, it only rationalizes why it's a worldbuilding question in the first place (the "world" is the AI's world).

Invitation

Therefore, I would like to invite WB participants to reopen the question. If you feel strongly that the question should not be reopened, please give a clear description as to why it should not be reopened.

Please note that before closure two answers were posted by people who understand what software testing is.

  • 1
    Agreed. I was one of the closers, but now I'm voting to reopen. – Renan Sep 5 at 17:11
  • 1
    What if we're of the opinion that it should be re-opened because of clarification, but then closed for being broad or opinion-based? – Frostfyre Sep 5 at 19:30
  • 1
    @Frostfyre, that's perfectly fair. However, I'd personally challenge opinion-based due to the provided criteria for a best answer. My hope is since so few participated in the Sandbox effort that such votes would include an explanation of why the vote was cast. Great Duck really is trying to understand the issues. – JBH Sep 5 at 20:02
  • 1
    I don't like the assertion that "people must be voting to close because they don't understand it", or the rest of that paragraph in this question. It seems unnecessary and, quite frankly, rude. – Kamil Drakari Sep 5 at 20:25
  • @KamilDrakari, how does one explain an uncomfortable observation without offending someone? The only way to justify that my use of the phrase was wrong would be to prove that its underlying assertion is wrong (which I'm open to). Welcome to the fundamental problem (and benefit) of free speech. Your being offended is the worst possible reason to change what I said, especially when that change would weaken the very point I was trying to make. – JBH Sep 5 at 20:33
  • @JBH I don't have "proof" one way or another that the assertion is correct, and neither do you. The first sentence in that paragraph could have "I think" added to the front and be at least reasonable and productive. The rest of that paragraph does not strengthen your point at all, it's just an insult. – Kamil Drakari Sep 5 at 21:22
  • I should point out that if I had asked why we would start a war in a simulation in general then RonJohn's comment would make sense, but I agree @KamilDrakari that the comment itself is based on the assumption that I am asking why a war simulation would be done in general. Can you yourself give a reasonable logical argument why that comment applies to the question of what benefits a war test would provide in determining software stability? If so then I will gladly concede but based on that comment at least one close vote was in error. – The Great Duck Sep 5 at 23:43
  • @KamilDrakari If pointing out a logical error is wrong then half of Math Stack Exchange answers are in violation of policies for pointing out the logical flaws of a user's flawed reasoning in their faulty proof or incorrect solution to a problem. Surely we are civil enough to allow similar acts of simple debate to be warranted in our posts on all of SE? – The Great Duck Sep 5 at 23:45
  • I also made sure to nominate my own question for reopening as I only just now got home and saw the closure. I hope it is obvious that I don't support my own questions closure unless a valid argument can be provided that I objectively agree with. I personally do not like closure reasons of "it's unclear". It's a particularly annoying close reason imo. I feel like it's used very often as a dumping ground for "tl;dr" which is annoying when generally the opposite of a vague question is to add more content. However that can be discussed later on in another meta thread. – The Great Duck Sep 5 at 23:54
  • Deleted my earlier comment re keeping closed: The Great Duck's response below comes very close to clinching the deal for me! – elemtilas Sep 6 at 2:15

As the original asker of the question I can only be neutral on the issue of closure as I am naturally biased towards it being open and I obviously cannot argue against closure if an actual reasonable argument is given for it. Instead I will bring up some points on the question that were somewhat lost in the question. I think some of these statements were removed because it made it either too story focused or were irrelevant. However, others might be able to reincorporate them better than me and JBH were able to, because as it stands I would lying if I said the original question has not been effectively replaced by a broader version. This could be good or bad, and this not out of malice. I am not saying the edits made were wrong, but that perhaps they can be rethought to further improve the question.

  • The original question stated that the developer was a 20 year old developer who discovered AI by accident and built a simulation like a video game where all the NPCs are AI and that this was primarily the result of general homebrew tinkering. In other words the motivation for building it could be argued as either curiosity or an attempt to build some kind of software that would be distributed as a video game.

  • The original question was looking for arguments that would convince a scientist inside the simulation leading one of the larger more peaceful countries to actually initiate the war. The argument had to be able to convince them through scientific reasoning however as I myself realized and other pointed out this AI would be built by the developer therefore any sufficient argument to convince the developer of its rigor would be sufficient to convince the scientist. Anything otherwise would imply that the AI knows more than the knowledge given it. External tools or the computer determining things from facts that are otherwise not human known could account for this, but the point was that the scientist was unnecessary.

  • the original question didn't state this outright but real world physics don't have to apply. In an extreme example every inhabitant could be a 2D texture variant of mario or any other enemy or NPC from the original SMB where they communicate with sounds that just... emanate from them. Or for that matter a thorough recreation of the world of flatland. After all, 2D games are all the rage and it would be interesting enough to conceive someone in their basement cooking up for homebrew.

  • The developer is an amateur at general software development. They might be incredibly skilled at programming (well obviously they have to be cause they invented AI) but they have no formal experience with actual software engineering and the formal process in development. Whether or not this factors into an answer based on a misconception of some kind or not is up to the discretion of answerers.

  • I was primarily looking for answers based on existing concepts most likely stemming from video game development testing likely relatable to the testing of MMORPG systems as this simulation is effectively a variant of an MMORPG where every player is an AI. Other answers are acceptable but to further stress, the idea here is that the simulation doesn't have to abide by any decent notion of real world physics. It just has to have a logical realistic notion of an open world that one could compare to some kind of realm in a video game.

  • Magic exists in this simulated world. period. It exists. The more general variant of the rewritten question might not need this detail but it certainly serves as a glaring example of why RonJohn's comment of governments making it to test cold war scenarios doesn't apply.

  • Thank you for this!! This does actually clear many things up! What I would like you to do, as the OP, is go back to the Main question and edit in these things that bring out the worldbuilding nature of your query! I'd also like it if you edited the actual question. You actually all but ask the question in bullet point 2: (what) would convince a scientist inside the simulation leading one of the larger more peaceful countries to actually initiate the war? Just indicate that you're looking for an external, software testing means of triggering the AI leader to act! – elemtilas Sep 6 at 2:12

I looked at the question at 2300 ET, 5 Sep 2018

I didn't understand it. It asks about quality analysis tests, but I can't clearly see what the objective of the software program is. I am a software engineer. I write testing suites all the time. But you have to know what you are testing for. A test is supposed to ensure that the software correctly does what it is supposed to do. What does the software do? Simulate a world war? Why do you even need to test that, it isn't real. You test software to make sure it conforms to the real world.

I really don't understand what this software is supposed to do, so I tossed it back in the review queue for closing. There is a lot of discussion about his topic, and normally I read everything and try to offer an opinion. But I'm drunk, so I didn't. Despite being drunk, let me offer my opinion: there may be a disconnect between people who do real live software engineering and testing and those who don't.

  • Looks like it's been magically reopened. The question itself is still off-topic. – elemtilas Sep 6 at 3:12

I did not vote to close this question, I just saw it now, but the question I have is what the role of the AI in the simulation would be.

Is the AI a soldier or general, or is its job to create a simulation of a war? I think if the OP were to answer that question, that would help clarify. Is the AI supposed to come up with a strategy for one or both sides or to be a creator who invents the world and the war in which everything takes place. If the AI plays the role of creator, how would an AI with human level intelligence simulate many other minds completely different from it's own?

  • I'm not sure you you came to these conclusions. For example, as said in the question, a computer can produce an "accurate simulation of large numbers of human minds." There isn't one AI. There's an entire world of AIs. An AI isn't the developer, the developer of the AIs is the developer. – JBH Sep 5 at 23:31
  • In the original question before edits the premise was that the simulation consisted of an entire population of AI NPCs essentially consisting a society or civilization. As for the purpose I would say that the purpose is never particularly made clear to the inhabitants of it. Considering the person making the simulation is just some 20 something year old mad genius operating out of their basement the end user could be people who play video games, anyone interested in it because it would be THE discovery of such AI, or anyone feeling big headed enough to literally wish to play God. – The Great Duck Sep 5 at 23:31
  • @JBH I think he simply means that the end goal of it might factor into testing which is true, but I think I would say that since the software is being uploaded to the internet and is efficient enough to be run on maybe even your computer I would say that it's probably best marketed as either a video game or just some really interesting freeware. – The Great Duck Sep 5 at 23:34
  • @TheGreatDuck Got it, you might want to add that back into the question, it seems like an important price of information. – John Locke Sep 6 at 10:31

You must log in to answer this question.

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