Posts for Warp


Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Zarmakuizz wrote:
Welp, they are now playing Emerald, no ROM hack, no domecracy at all.
It will be turned on by the streamer if they get stuck long enough, though.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
For a moment I thought that "For Whom The Sweetie Belle Toils" would be cringeworthy or at least boring, but in the end I thought it was pretty sweet. Somehow especially Rarity's characterization felt perfect and endearing (much unlike her behavior eg. in "Sleepless In Ponyville".)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
ais523 wrote:
So in other words: this run is too similar to the standard runs for me to start with
I don't like at all how this run completes the game, but I have to object to that particular point. While the beginning is obviously the same as other runs, this is pretty much by necessity. However, by the end it becomes completely different (completely different route, different techniques), so I think this particular part of your objection is not valid. However, I like your notion that the run jumping to an end routine via arbitrary code execution isn't actually completing the game. I think there's merit in that idea. (It's a big like if arbitrary code execution were to be used to show the "game over" screen of a game. Ok, it showed the screen, but it didn't really complete the game, did it?)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Dessyreqt wrote:
http://tasvideos.org/2027M.html That movie has a star and not only uses the above "debug code" but also uses a code at the beginning while on "Cheat Mode" to reconfigure the layout of the levels. Interestingly, that run had no controversy about acceptance. It was faster than the existing run of the given criteria, and so it was accepted. Why would this run be any different?
As the old adage goes, past mistakes don't justify repeating them. (I'm not saying accepting that run was a mistake. I'm just saying that arguing that this should be accepted because that one was too is not valid. In general, runs don't set precedents.)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
I have definitely seen more boring TASes. Perhaps not the most entertaining TAS in existence, but pretty entertaining nevertheless.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Patashu wrote:
Please answer this: Why is an Earthbound TAS that accesses the debug menu unexpectedly OK, but a Super Metroid TAS that accesses a debug feature unexpectly NOT OK?
For what it's worth, IIRC I opposed and voted against the idea. I also remember there having been a discussion (although I don't remember if it was within that thread or somewhere else) that using a cheat code (like the konami code) directly is not ok, but triggering the exact same cheat routine via glitching is ok. I found that argument nonsensical back then, and I still find it nonsensical today.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
amaurea wrote:
The Vault was supposed to solve the "problem" of too many categories by hiding TASes that aren't deemed entertaining enough.
Actually I think the idea with the vault tier was more like "every game deserves a WR completion, completely regardless of entertainment factors". As to not make the tier go completely out of hand, it was decided that the two "WR" categories are "anything goes" and "100% completion".
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
I have the feeling that the OP is being genuine, rather than wanting this thread to be a joke, but people are not taking it seriously and making fun of it instead. In a rather inappropriate way, I might add.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Being a party pooper again, but... I was super-excited when I saw this submission... until I saw the three dreaded words: "arbitrary code execution". I'm just not in the camp who gets all excited by a game being completed by glitching into running arbitrary code, and then jumping to some arbitrary end routine. I understand perfectly that others are, but it's just not my cup of tea. I just don't like the idea. Kudos for the technical achievement (and I do honestly appreciate the work that went into this, and has overall been put into Super Metroid speedrunning in general), but this kind of completion just isn't for me. (Since I have conflicting views on what the voting should mean, and what it says, and I'm not technically speaking opposed to publication, I'll just resolve the conundrum by abstaining from voting. It's not a statement or a protest, it's just personal preference.)
Post subject: Re: Seemingly everyone hates c++
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Tub wrote:
To "learn more", C or even assembler will certainly force you to help you to learn more about the inner workings of your computer. Which is a good skill to have if you ever need to debug a binary third-party-dll, an emulator rom or similar. Understanding the basics is also important if you ever need to really really optimize some algorithms.
We have already been through this. A beginner doesn't need to "debug a binary third-party dll, an emulator rom or similar". Heck, I have been programming professionally for over 10 years and I don't need to do that. You don't need to know C or asm to learn how the CPU works and what kind of code a compiler produces. You don't need to know C or asm to "really really optimize some algorithms". In fact, you extremely rarely, if ever, need hardware-level optimization. Even in the extremely rare cases that you do, you don't need to know asm for this. Asm doesn't help you at all. What you need to know is how things like CPU caches and memory fetches work. Neither C nor asm teach you this in any way, shape or form. (Yes, I'm serious. It's not like machine code dictates, or even shows in any way, how memory caches work, how large they are, and how data is fetched from RAM to caches.) I don't even understand why this myth is so persistent. Here is an example of an article I once wrote about a low-level optimization. Do you know how much C or asm you need to know to understand how to make this kind of optimization? Zero. All you need to know is how memory caching works, and how your particular programming language arranges its data containers in memory. There is no need to know even a single machine code instruction.
Post subject: Re: Seemingly everyone hates c++
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
antd wrote:
Everyone says C. But I wonder if it is used for anything beyond device drivers and such anymore. And those same people tell me that they hate C++.
I would not recommend C for a beginner (or for anybody for that matter.) It may be easy to write your first "hello world" program, but immediately when you need to do anything requiring a dynamic amount of memory allocation, you run into difficulties. C++ makes many things a lot easier, but it requires you to know what to use (to make those things easier) and what not to use, and how to properly use them. Unfortunately there aren't many good C++ tutorials out there. Other, "safer" languages may be better for beginners. Most of them will offer most of the tools you need on the get-go, and most of them will give runtime messages when you do something wrong (such as accessing a data container out of boundaries). If you want a language you can write a GUI'd program with easily, I suppose C# could be one possibility, but it will be Windows-only. (Theoretically the language is available for other systems, but in practice you can forget about it.) (Personally I program in C++ and Objective-C++ as my payjob. It's all the rage in the iOS side of the world, and that suits me just perfectly. However, it does require quite a lot of experience to get fluent at. Probably not something a beginner can immediately get into.)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
I tested what happens if I only mark the twin primes in the ulam spiral. The result was rather interesting (if not even hypnotic.)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
I know this isn't in any way a new idea, but just for the fun of it I made a small program that draws the so-called ulam spiral, but with composite numbers colored as well. Each pixel is colored with a shade of gray depending on how many terms there are in its prime factorization. The shade of gray is 255/n, where n is the amount of prime factors. (Ie. it's 255 for primes, 127 for numbers with 2 factors, 85 for numbers with 3 factors and so on.) In other words, the darker the pixel, the more prime factors it has. Besides the traditional prime (ie. white) diagonal patterns, I find it interesting that there are some relatively wide diagonal dark areas. These seem to be diagonals where there are many numbers with a significant amount of prime factors. Another interesting detail is that this coloring also emphasizes many horizontal and vertical lines with lots of primes in them (and which are not so evident in an image where only the primes are drawn.)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Honestly, Maud has become one of my favorite characters of the entire series.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Compared to the completely underwhelming, even annoying, previous episode, this episode (Maud Pie) was a breath of fresh air. In fact, IMO it's one of the better episodes of the series by far.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Once again theory and practice do not meet: In theory this should be one of the most boring runs ever. It's the exact same move done again and again on every round of every single stage, with not much else happening. There's very little to see here. In practice, however, I found this (once again) very interesting to watch. It's actually amusing to see how each opponent is ohko'ed as fast as possible and the game completed so incredibly fast. Bonus points for having to do subtle things at the beginning of some fights to avoid a hit from the opponent. I actually found that interesting. Once again this divides opinions. I'm firmly in the camp that this was not only entertaining, but in my opinion so much that it deserves a moon.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
thatguy wrote:
The sad thing about this submission is that, if FractalFusion had just used the code execution to finish the game rather than promising a payload that didn't really materialise, then its reception would probably have been quite good.
That's because its goal would have then been different, and thus would be subject to different standards. Personally, however, I'm not very happy with arbitrary execution hacks jumping to the end of the game. The reason for this is that once you can freely decide where to jump in the game's code, the definition of what constitutes "jumping to the end of the game" becomes pretty much arbitrary. How do you define the "end of the game" in this respect? Exactly which part of the game code can be considered the "end of the game"? Is it right after the final boss? The end of the possible epilogue? The start of the end credits? The end of the end credits? Something else? Why that particular point, and not a point earlier or later? It becomes a question of subjective opinion, and thus it's pretty arbitrary and up to the person deciding. Personally I would prefer if the game would be allowed to naturally reach its endpoint (end boss is defeated, and the game proceeds normally from there.) And if the question is "well, could you thus jump to the end boss via arbitrary code execution? Who decides where it's ok to jump?" Well, precisely: Jumping to an arbitrary execution point is subjective. The solution is pretty simple: Just don't do it. But that's just me (and I understand I seem to be an extremely small minority on this.)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
You could have literally made the game do almost anything, and you choose to make it just print some numbers on screen. Definite no vote. This type of run must leave the viewer in awe. Like that SMW run that allows you to play arcanoid and snake, which elicited awe and applause in the spectators. It has to be something marvelous. To obsolete a previous run of this category, the new run has to be even more awesome than the previous one. This was not. The technical achievement to get the arbitrary code execution faster is commendable, but the payoff was not worth it. With a much better payoff it would have probably been a different story. (And please, please no more "pi" nonsense. The previous "pi" submission was boring and nonsensical, and personally I objected to it obsoleting the previous run. Do something much more visual, much more awesome. I'm still hoping for someone to make that "rickroll" version. That would be something worthwhile.)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
feos wrote:
You're talking about this run, right? If so, why not call it playaround? (Just to put the reason clear.)
Well, if the run aims for the best score possible (which I think it does), then why not use that as a branch name? It's a pretty well-established branch name and is more accurate.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
A CPU being "8-bit" does not mean it's unable to make calculations with anything that exceeds 255. Anyway, even if a game would count how many times it has been completed, and even if it did so with one byte, all that would happen is that the counter wraps back to zero. It then depends solely on whether the value zero is ok for it or not. (Assuming that it starts at one, and the game was not programmed expecting it to have a value of zero, something might happen. Although that something is most probably just some value somewhere showing zero instead of whatever it's supposed to show, if the game shows it in any way at all. For the zero to actually cause glitching would require quite peculiar programming decisions...)
Post subject: Re: Kill screens and NES
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
andrewg wrote:
If you beat an NES game 256 times in a row, do you think there could be strange outcomes or kill screen like pacman and others?
Why would there be? How many games do you know of that count the number of completions (and does so with a single byte)? From them, how many do you know that will glitch when the count wraps to 0?
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
For the game to be playable by thousands of people it needs these characteristics: - No precise movement required to progress. 4 directions is actually pretty optimal. - One single movement command ought to make significant progress. Grid-based movement like in old JRPG's is pretty ideal for this. - Pretty much has to be turn-based. It must not have anything that's based on reaction time. Again, old JPRG's fit the bill nicely. - No possibility of getting stuck forever, no possibility of making the game unwinnable. (Pokemon, and basically all other JRPG's, make sure of this by making it impossible to permanently toss or use out items that are crucial to beat the game. (*) ) - No gameovers or limited lives that make you, when they run out, start from the beginning of the game. (*) Although incidentally in the 1st generation pokemon, possibly in others as well, you can get completely stuck even without abusing any bugs. However, this requires a very precise set of actions and circumstances that are extremely unlikely to be met even with such randomized input.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
I don't understand why speed couldn't be a perfectly reasonable goal. Not aiming at the perfect score would strongly feel like sloppy play. If a hole can be done in 2 shots, and the player does 4 just to do something "funny", it will feel unoptimized, sloppy and a waste of time. It should be done in 2, if that's the minimum amount that's physically possible. How to do those 2 shots is another question. It could go way off course than then from there make a miracle shot to the hole. In the vast majority of cases, however, that will probably not be the possible, and a completely optimal (also timewise) play will probably be as good as anything else. So I'd say: Primary goal, perfect score. Secondary goal, don't waste any time unnecessarily. Branch name "max score" unless the run really aims at the shortest possible time (in which case the branch would be unnamed.)
DRybes wrote:
This game doesn't need a branch, because we will only ever have one run of this game.
Because the new policy that's under consideration/adoption says that an unnamed branch means "any% fastest possible completion". Any other goal should have a branch name, even if there's only one run of a game. Making a publication with no branch name that nevertheless is not the fastest possible completion would break that naming convention and would be confusing.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
I agree. The purpose of TASing is to show what a superhuman playing a videogame would look like. That usually means mopping the floor with the game on the hardest difficulty setting. Using the easiest setting kind of makes the whole point moot (unless there's a very good reason to use it.)
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
henke37 wrote:
Uhm, it's a minimum score run. Golf is about having a low score.
I think it's a question of perspective. To "maximize the score" you try to go as much "under par" as possible. The more under par you go, the better. "20 under par" is better than "10 under par". In that sense, you are maximizing your score. Nitpicking aside, I'd say it would be a good idea to keep branch names as consistent as possible and sensible even between games.