Gamebreaking bug for worldspaces bigger than 4 quads in widt : V - Skyrim

» Tue Jun 19, 2012 2:48 am

Yes they could easily not bother, but there are a couple of reasons to think they will take it seriously enough to fix it.

1. Leaving it as is will affect the longevity of interest in the game. New worlds attract new players = continuous income.
2. Not bother would for at least some people make this the last game in the series.

Thats not griping or moaning, just plain honesty. I wouldnt buy another, would you? Who here would if they suspected from the word "go" that their pet project such as Merp, Skyrim extended, other realms, other provinces, getting Morrowind into Skyrim, Mesogea etc etc was unlikely to happen.

Personally I prefer to think from past experience. We had a LOD bug in Oblivion, where you could make lands as big as you wanted (up to 16x16 quads) but beyond 8x8 quads the LOD would not show up so you needed permanent fog or floating land. When S.I came out and the CS was upgraded, the problem was fixed. The large world mods picked up speed and drew in the punters. Punters make profit.

Hopefully with over 5000 views this thread may start gettig a bit of attention from devs?

» Mon Jun 18, 2012 9:06 pm

Hopefully. Its on the critical list in the bug thread. Has anyone had a reply from a dev yet, or from anywhere. Any comments on twitter?

» Mon Jun 18, 2012 1:45 pm

I think they only let us know what they are fixing if there's a full community uproar about it, like the NavMesh bug... I personally don't expect they'll tell us in advance if they're going to fix it, but I am going to keep making this bug known as long as we don't.

» Mon Jun 18, 2012 6:47 pm

Any progress on putting up calls for tweets? There hasn't been a response on twitter yet but we might get something if they get enough activity there.

» Mon Jun 18, 2012 11:27 am

The issue with this bug is that it wont create uproar until people reach the stage of hitting it. In Oblivion it took a fair bit of time for new worlds to develop and for there to be a massive desire for them. Dune, MERP, Anequina, Mesogea and the other province mods took time to develop and to build up. It was months before we found out about the distant land bug beyond 96 or 128 cells (I think it was the latter) and it was only because of SI coming out some months later that I believe we had it fixed. By the time the error had been found we had quite a few projects on the go and a large portion of the modding and playing community wanting them.

Here we are suffering from foresight. We knew there were issues in Oblivion about using worldspaces larger than 8x8 quads so we went straight for the jugular and bang! found errors within days of the CK being released, long before most of the community were interested in large worldspace and extended worldspace mods. If this bug had been hit several months down the line, with large numbers of people again working on MERP, Anequina, Dune, Skyrim extended, other provinces, Mesogea, custom worlds of all types etc, then I think right now we would have a massive uproar. It is only because the navmesh issue is fast and easily found we have thankfully got the current furore.

» Mon Jun 18, 2012 4:36 pm

Bethblog tweet advises new beta update on steam

http://www.bethblog.com/2012/02/20/new-beta-update-available-on-steam/

Says "Today on Steam we’ve uploaded a new Beta Update (1.4.26) highlighted by some tweaks modders have requested for the Skyrim Workshop."

Just "tweaks" then - no mention of fixing the enormous great bugs for either navmesh or large worldspaces

» Mon Jun 18, 2012 2:52 pm

Only for the Workshop it says...

» Mon Jun 18, 2012 4:55 pm

The issue with this bug is that it wont create uproar until people reach the stage of hitting it. In Oblivion it took a fair bit of time for new worlds to develop and for there to be a massive desire for them. Dune, MERP, Anequina, Mesogea and the other province mods took time to develop and to build up. It was months before we found out about the distant land bug beyond 96 or 128 cells (I think it was the latter) and it was only because of SI coming out some months later that I believe we had it fixed. By the time the error had been found we had quite a few projects on the go and a large portion of the modding and playing community wanting them.

Here we are suffering from foresight. We knew there were issues in Oblivion about using worldspaces larger than 8x8 quads so we went straight for the jugular and bang! found errors within days of the CK being released, long before most of the community were interested in large worldspace and extended worldspace mods. If this bug had been hit several months down the line, with large numbers of people again working on MERP, Anequina, Dune, Skyrim extended, other provinces, Mesogea, custom worlds of all types etc, then I think right now we would have a massive uproar. It is only because the navmesh issue is fast and easily found we have thankfully got the current furore.

That's also because the navmesh-issue is so critical, that even beth must responde and fix it immeadetly. And since they also have the workshop since skyrim, they need to make more updates.

But this issue of the worldspace-bug is also very important for modders who want to create large custom worlds, and this should be fixed immeadetly with or after the navmesh-bug.

» Tue Jun 19, 2012 12:16 am

That was my point. The Navmesh bug was so immediately because its easy to find and catastrophic across the board. Heightmapping is a bit more specialist and would ordinarily have taken a lot longer to find. Once found though, the outcome is the same, lots of people hacked off.

» Mon Jun 18, 2012 5:25 pm

Bump.

» Mon Jun 18, 2012 1:31 pm

This is a bug that really needs to be fixed. Half the reason I buy Elder Scrolls games is for the mods and mostly the big mods that expand the world. Its the reason I still play Morrowind. With mods like Tamriel Rebuilt still being developed it keeps the game fresh.

If I find some extra time over the next couple weeks I may look into this to see about finding a viable workaround or fix. Unfortunately I will be too busy at least until next Tuesday, hopefully after that I can look into it. Although it would be nice if Bethesda just fixed it...

» Mon Jun 18, 2012 5:50 pm

Thats massively appreciated.

» Mon Jun 18, 2012 10:12 pm

Perhaps we can figure out why it doesn't happen to the player character and non-actor dynamic forms. We could first see if it's perhaps a physics setting inside the NIF file, like if perhaps the bug can be disabled by giving a weight of 0 to the NIF model. Turning off certain types of AI should probably be tested as well, perhaps it is related to checking for idle animations or something like that (the player character doesn't use those).

I'm sure it is not merely a Havok Physics bug, because only non-player actors are affected. There is another variable affecting this. Narrowing down that variable might be valuable.

» Mon Jun 18, 2012 8:04 pm

Not sure how much it helps or not, but in a http://www.youtube.com/watch?v=qndePb5jEO4, if you pay really close attention during the last few frames, when the console is open and he's getting the player's X position. The 262,144 value where it suddenly begins happening. That falls right on a power of 2 boundary, which is always of interest when dealing with computer programming.

» Tue Jun 19, 2012 2:23 am

Not sure how much it helps or not, but in a http://www.youtube.com/watch?v=qndePb5jEO4, if you pay really close attention during the last few frames, when the console is open and he's getting the player's X position. The 262,144 value where it suddenly begins happening. That falls right on a power of 2 boundary, which is always of interest when dealing with computer programming.

"He" is Maegfaer, it's his youtube channel i believe, just thought I'd clear that up..

» Mon Jun 18, 2012 5:08 pm

If the issue is the x-coordinate then the problem shouldn't be happening at the same cell coordinate for positive and negative values, should it?

Cell 0,0 has coordinate (0,0) at the bottom left so if it happens at the x=64 to x=65 cell boundary then for positive x:

Cell 64 is a power of 2 and there are 4096 x/y units per cell, also a power of 2. So the +x cutoff will be a power of 2: 64 x 4096 = 262,144 at the transition to x=+65

But -ve x is effectively numbered 1 higher - the cell to the east of the origin is cell (0,0) but the cell to the west is (-1,0) - so you reach (-64,0) having travelled only 63 cells west of the origin.

So if the cutoff is an x value of +/- 262144 then the cell boundaries where it happens will be the transitions from x=+64 to x=+65 and from and x=-65 to x=-66

So if the effect kicks in at the x=64 to x=65 cell boundary and the x=-64 to x=-65 boundary then the effect is cell numbering based

But if the effect kicks in at the x=64 to x=65 cell boundary and the x=-65 to x=-66 boundary then the effect is position based

Or have I got my brain twisted again?

» Mon Jun 18, 2012 8:18 pm

I did some testing and I believe to know what is causing this. It's the water. (Or we have a second bug)

I added a line at the border between -65,0 and -64,0 and I set the waterheight below the heightmap (in both cells). I also added some creatures.
Ingame the waterheight is correct in cell -64,0, but if I go too cell -65,0 I am under water (what I shouldn't be, because both cells got the same waterheight). I think the engine has some problems with the water in cells with -64