Author Topic: Experimentation on breaking SSF  (Read 14757 times)

Katharine Berry

  • Guest
Re: Experimentation on breaking SSF
« Reply #15 on: August 15, 2007, 06:26:08 PM »
That's the same general effect you'd get from mad physics stuff (which was probably the cause) - the number marked "Time Dilation" shows just that - 1.00 means that time (specifically the physics engine) is running at 1x normal speed. If that gets below 0.95 it'll be noticable, below 0.01 and the sim's dead. And you can't move (or, rather, you can, but do so really slowly).

Furthermore, high levels of physics engine activity may use up CPU cycles that could otherwise be spent monitoring network connections or accepting new ones, essentially logging everyone out and making it impossible to log in again until the sim detects it and restarts (unlikely by this stage), or Linden Lab manually intervene.

Offline Marko Schomer

  • Moderator
  • Hero Member
  • *****
  • Posts: 631
    • View Profile
Re: Experimentation on breaking SSF
« Reply #16 on: August 15, 2007, 06:52:21 PM »
I stopped experimentation after the initial warning about the possibility of crashing. However, I'm not actually sure about the risk involved. Since the objects move apart immediately, and there's nothing to prevent them doing so fully, I would have thought the probability of serious lag would be extremely low.

I'm also unclear on what theories one can form about the act of throwing objects in odd directions.

My original investigations were into how strong the barrier against intersection was (and whether it was based on velocity, momentum, or other factors) Since there's evidently no such force in real life, it's quite interesting how one works in SL (the fact that it's not an absolute barrier is interesting in itself).

Offline Explo Schomer

  • Moderator
  • Hero Member
  • *****
  • Posts: 2386
  • Nice, eh?
    • View Profile
Re: Experimentation on breaking SSF
« Reply #17 on: August 16, 2007, 01:51:22 PM »
Quote
That's the same general effect you'd get from mad physics stuff (which was probably the cause)

I believe the actual cause was the generation of thousands of little wooden cubes at once. If it was due to physics experiments you would notice it once, and then once the island came back online it would happen again slower, to work out what happened. ;)
'I am the gadfly'-or at least, I'd like to be

Question everything, including this.

Katharine Berry

  • Guest
Re: Experimentation on breaking SSF
« Reply #18 on: August 16, 2007, 02:34:02 PM »
Quote
That's the same general effect you'd get from mad physics stuff (which was probably the cause)

I believe the actual cause was the generation of thousands of little wooden cubes at once. If it was due to physics experiments you would notice it once, and then once the island came back online it would happen again slower, to work out what happened. ;)

That very rarely crashes a sim (unless the cubes were physical), although it tends to bring down the entire grid (not hard to do, is it? :P), which, if LL knows who did it, will generally result in a permaban and being reported to the appropiate authorities for causing a denial-of-service attack - LL can fairly easily demonstrate significant losses from a few hours of downtime or unusable-service-time. If this is done on a PI repeatedly, the sim is liable to be confiscated due to the sim owner's negligence.

Since there's evidently no such force in real life, it's quite interesting how one works in SL (the fact that it's not an absolute barrier is interesting in itself).

It is an absolute barrier - it's seen as part of the object (as far as the physics engine is concerned, all objects are slightly larger than they appear). However, SL is notoriously bad at actually preventing intersecting objects, or objects passing right through eachother.

I stopped experimentation after the initial warning about the possibility of crashing. However, I'm not actually sure about the risk involved. Since the objects move apart immediately, and there's nothing to prevent them doing so fully, I would have thought the probability of serious lag would be extremely low.

It seems I misunderstood what you were doing, as that wouldn't cause any issue. But I maintain that meddling with the physics engine is a bad idea anyway.

Offline Marko Schomer

  • Moderator
  • Hero Member
  • *****
  • Posts: 631
    • View Profile
Re: Experimentation on breaking SSF
« Reply #19 on: August 16, 2007, 06:03:57 PM »
Quote
However, SL is notoriously bad at actually preventing intersecting objects, or objects passing right through eachother.

Which, in effect, means that the barrier isn't absolute. :P

Quote
It seems I misunderstood what you were doing, as that wouldn't cause any issue.

Thanks for warning me anyway. I'd far rather miss a few experiments than crash a grid and risk Schome suffering for it. As to meddling with the physics engine, I think as long as the risk is very low then it's worth doing experiments. Otherwise what can be done is greatly limited, and you get a situation similar to that faced with health and safety in RL.

Katharine Berry

  • Guest
Re: Experimentation on breaking SSF
« Reply #20 on: August 16, 2007, 09:08:14 PM »
Yeah, but here bad stuff actually happens, unlike Health and Safety. :P

I say this having crashed my own sim more than once.

Quote
Which, in effect, means that the barrier isn't absolute.

It's just as absolute as the object, and when that new physics engine (which I now get to discuss the implementation of with LL, having signed an NDA to say I can't tell you) appears, all this will stop happening anyway. :P

Offline Dan

  • Hero Member
  • *****
  • Posts: 1242
    • View Profile
    • Dan's Blog
Re: Experimentation on breaking SSF
« Reply #21 on: August 16, 2007, 09:41:18 PM »
Are you saying they're actually going to put in place a quality games-standard physics engine?
SL: Woop Superior (MG) / Woop Kamachi (TG)
RL: Dan

Katharine Berry

  • Guest
Re: Experimentation on breaking SSF
« Reply #22 on: August 16, 2007, 11:12:01 PM »
I am. :P

Offline Dan

  • Hero Member
  • *****
  • Posts: 1242
    • View Profile
    • Dan's Blog
Re: Experimentation on breaking SSF
« Reply #23 on: August 17, 2007, 07:46:48 AM »
Thank goodness.  This has been one of the things that has long bothered me about SL.
SL: Woop Superior (MG) / Woop Kamachi (TG)
RL: Dan

Offline mgaved

  • Hero Member
  • *****
  • Posts: 1164
    • View Profile
Re: Experimentation on breaking SSF
« Reply #24 on: August 17, 2007, 08:06:46 AM »
Send forth details as you get them.....  :)
Teen Second Life: Mark Sparker / Second Life: Redmark Hax

Offline Dan

  • Hero Member
  • *****
  • Posts: 1242
    • View Profile
    • Dan's Blog
Re: Experimentation on breaking SSF
« Reply #25 on: August 17, 2007, 08:12:59 AM »
I think with an NDA in place that's going to be difficult  :(  Same when I'm beta testing games, I want to put out details on my blog but can't until we're near the end and the NDA gets lifted.
SL: Woop Superior (MG) / Woop Kamachi (TG)
RL: Dan

Offline Achilles Schomer

  • The Hawaiian Shirts
  • Hero Member
  • ***
  • Posts: 1687
    • View Profile
Re: Experimentation on breaking SSF
« Reply #26 on: August 17, 2007, 10:29:45 AM »
a assume a NDA is a secrecy directive?
My posts are like buses...you wait days for a reply then they all come at once!

Offline Marko Schomer

  • Moderator
  • Hero Member
  • *****
  • Posts: 631
    • View Profile
Re: Experimentation on breaking SSF
« Reply #27 on: August 17, 2007, 10:36:58 AM »
Yeah, but here bad stuff actually happens, unlike Health and Safety. :P

It's quite odd when worse things happen in a virtual world than in the real one. :P

Quote
when that new physics engine (which I now get to discuss the implementation of with LL, having signed an NDA to say I can't tell you) appears, all this will stop happening anyway. :P

Really? I thought Havok 2.0 would just reduce problems with physics, rather than eliminate them. If not, then the physics group will have to check current laws all over again.  :(

Offline Dan

  • Hero Member
  • *****
  • Posts: 1242
    • View Profile
    • Dan's Blog
Re: Experimentation on breaking SSF
« Reply #28 on: August 17, 2007, 11:49:12 AM »
An NDA is a non-disclosure agreement.  Common policy when getting good people to evaluate/contribute to some kind of development to sign one of these when they aren't directly employed by the company.  Thus in the case of Katharine and the physics engine or myself and the likes of Lord of the Rings online we would be chatting to working with the developers and/or testing the latest developments, but having signed a non-disclosure agreement cannot publicise any of the information we have.
SL: Woop Superior (MG) / Woop Kamachi (TG)
RL: Dan

Katharine Berry

  • Guest
Re: Experimentation on breaking SSF
« Reply #29 on: August 17, 2007, 12:29:53 PM »
Really? I thought Havok 2.0 would just reduce problems with physics, rather than eliminate them. If not, then the physics group will have to check current laws all over again.  :(

Havok 2 was abandoned in 2006. New physics project. :P