Fix Your Timestep: Crazy Internet Stalker!
A few weeks ago, this guy called abaraba reads my Fix your Timestep article then posts a weird comment: “your technique has a flaw, blah blah blah (weird explanation that makes no sense)”. I reply that it actually, no it doesn’t and explain how he misinterpreted the article then this guy flips out and for the next few days posts these strange rambling paranoid comments about his new time-stepping technique all over my website.
Initially I try to reason with the guy and help him out, but I give up after it’s clear he just isn’t really interested or capable of learning. Delete 10 pages or so of rambling comments I forget about it and move on.
A few weeks later, I notice all these forum link referrals in my wordpress stats and I click through. Turns out after I deleted his comments the guy was so pissed off that he spent the last few weeks on some bizarre internet forum crusade to “get the truth out” about time stepping!
Check out the forum threads, they’re hilarious!
devmaster.net
intel.com
nvidia.com
bautforum.com
gamedev.net (#1)
This guy is fucking insane!
Ok, *cough* where was I?
Alright. He’s not just batshit crazy, he’s also completely wrong.
Dere is my quick rebuttal of abaraba’s timestepping technique:
- He attempts to solve the problem of “spiral of death” by decoupling sim time from actual time elapsed, which in the context of a game simulation would result in simulation slowdown as perceived by the player when under high CPU load. While this may be reasonable for non-interactive simulations, for games it is very undesirable.
- Furthermore the technique breaks multiplayer games, since each computer involved in the game simulation would advance time forward at different rates, according to simulation load, making it impossible to keep the two game simulations in sync.
- Modern physics simulations (eg. Havok) support iterative methods so that a lower quality simulation result may be obtained if the simulation runs out of frame time, which helps in cases where the game is simulation bound, a fact which abaraba seems to dismiss or is incapable of understanding (see intel thread)
- Finally, abaraba fails to understand the concept of temporal aliasing and the necessity for interpolation of render state when physics simulation steps at a fixed rate decoupled from render rate. Which is the whole bloody point of the fix your timestep article!
- Ironically, this results in his “improved” timestepping code actually being less smooth than fix your timestep!
In short, abaraba really has no idea what he is talking about, so take anything he says with a very large grain of salt.
UPDATE: Round Two – FIGHT!
devmaster.net
natscience.com
sci.physics.relativity
gamedev.net (1)
gamedev.net (2)
gamedev.net (3)
gafferongames.com (smackdown)
Better make that a whole rock of salt folks. Ayres Rock
I’ve been keeping an eye on him too since he started posting on the Intel/Havok forums. He’s also been on Bullet and Physics Forums (cached link):
http://www.bulletphysics.com/Bullet/phpBB3/viewtopic.php?f=4&t=2287
http://209.85.135.104/search?q=cache:pSrbxPRL-acJ:www.physicsforums.com/showthread.php%3Ft%3D265099&hl=en&ct=clnk&cd=1&gl=ie&client=firefox-a
Too bad you didn’t keep the archived comments around
-C
yeah they were pretty funny
check out this paper, i think it explains the whole situation pretty well!
http://www.apa.org/journals/features/psp7761121.pdf
Yeah, that idiot was an…interesting one to deal with on GameDev.
Well, for the record, I’ve found the whole thing hilarious.
I’m sure we’ve all met an “abaraba” or two in our time but he’s making himself a household name!
digg it up folks, only you can UNCOVER THE TRUTH about time-stepping!
http://digg.com/programming/The_Timestep_Crusader
An Aussie game developer in LA? If you’re ever in Austin TX, drop by for some Texas-style BBQ and/or tea and scones. I’m from Melbourne.
Awesome nigel, thanks!
I was watching as his first post on the Nvidia PhysX developer forums appeared, where I think the “crusade” originated. I don’t think he’s a troll- I think that he’s just very confused. I’m not going to go ahead and say that he has some kind of mental disorder, but if he does I wouldn’t be surprised. It’s just in the manner of how he rambles and talks in his posts, his thoughts are very unorganized and sporadic. Perhaps English isn’t his first language.
Well, seems you became famous and now have your very own stalker! Probably you will find his bedroom is covered with your pictures or your printed articles.
You guys should go easy on the poor guy, maybe he has Asperger syndrome (http://en.wikipedia.org/wiki/Asperger_syndrome). Could explain the intensity and peculiar way of speaking.
@Kris
*I* have Asperger’s Syndrome, and so do a lot of folks in the game industry. It doesn’t excuse his rudeness, only his passion.
go easy? you should see the comments he posts on my website
this guy had it coming
Hehe, so he has been here as well. I poked him around a lot at the nVidia forums. I was perhaps a bit childish as well, but I actually think he’s just very young, perhaps high school tops.
It was hilarious indeed. In the end I blocked his access to the Bullet forums and told him to go discuss his issues in other forums.
He did
I just feel bad for him. From reading his posts, it seems to me that maybe he doesn’t know English well enough to understand that he comes off as a big jerk. He also doesn’t seem to realize that he is receiving relevant critique of his ideas and presentation rather than personal insults. I’m also inclined to agree that there are probably some mental irregularities that he is struggling with, but I’m not really a psychologist…
Wow. He sounds just like the timecube crackpot. Maybe they should get together and create a new theory.
My theory is that the guy is some weird non-english speaking guy who runs the responses through babelfish before reading them, responds in his native language, translates that back to english, then posts it in the forum…
either that, or he is fucking batshit crazy
Don’t be surprised if he starts referring to you as a “Stupid evil bastard human”
. Actually he said something on the nvidia forums along the lines of: “you silly human person”. That’s kind of scary now that I think of it.
it all makes sense now, he’s not from this planet
Sounds to me like the guy might have a bit too much spare time on his hands. Perhaps he should consider a hobby!
Jess
http://www.web-privacy.pro.tc
I just came here looking for some help on setting up my physics time-stepping… Now here it is an hour later and I’m totally enthralled by the legend of Abaraba. Wasn’t expecting to get a great physics article AND a hilarious story.
Keep up the great work Glenn! I’ll be making a donation if/when I use your timestepping method.
Wow! I just wasted an hour of my life reading his garbage. Thanks Glenn!!
wow, i first read your comment as “reading this garbage” then I realized what you meant – *phew*
I looked through his stuff and it seems pretty uncomprehensible, also his manner of talking is very uncooperative, so I’m not suprised nobody listens to him.
But I was wondering about his points:
———————————————————————————————————–
Bug 1.) Few subSteps: “Moon gravity effects”
bUg 2.) A lot of subSteps: “Spiraling to death”
buG 3.) Interpolation VS. Temporal Distribution
Common mistakes and bugs related to fixedTimeStepping – UNCOVERED
4.) Design-Time Planning and Recognition of Hardware Min. Sys. Requirements
5.) “Scaling the World”: Gravity, Mass, Size and Hidden Effects of fixedTimeStepping
6.) Jitter, Wobbles, Choppiness and Slowdowns: Unstable Physics & Smooth Animation
———————————————————————————————————–
I would like to know if those are valid actual problems/bugs, and if they should matter to me if I am writing a physics simulator?
the key point to remember is to make sure that you are not simulation bound – it is quite obvious that if you want a realtime simulation at 60fps, then you actually need to be able to simulate your physics frame in less than 1/60th of a second
if you can do this it’s fine
i would say you want to use “fix your timestep” for only non-sim bound cases, eg. when render bound
if sim bound please use a separate render thread so you avoid the 1-2-1-1-2 split which means you need head room relative to render and sim thread as they go in and out of phase with each other
Hey, I noticed one of his comments is still on the “Integration Basics” article, the date is October 2008 if that helps to find it. I was trying to follow some of his threads when he was posting, talk about a headache-ina-box. Has anyone read his “book”? I stopped after the second paragraph and threw my hands in the air and said “Crap, I don’t have enough breadcrumbs to get home…”
Heh, I remember him from either gd.net or devmaster.net. Didn’t know about his crusade, tho. Oh, and thanks for the new color on my LCD: “coffee brown”.
I remember that guy. Met him on planet Dowgerong. Always been a crackpot, that fella.
ruh-roh looks like abaraba has discovered networked physics…
http://gafferongames.com/game-physics/networked-physics/#comment-11576
roflolwow thats absolutely over the top
This is insane! He’s been at it for over a year!
At one point he even said something like “I deserve a noble prize”.
He did manage to make me very upset, this guy is the future of trolling.
I’m getting away from here while I still can!
btw, I actually followed a link from StackOverflow – http://stackoverflow.com/questions/2009639/velocity-based-on-the-time-being-passed
not one of his forum posts.