Grey Goo Fence Stay or Go?

Pretty simple, the grey goo fence is breaking existing content. Should it stay and everyone fight over how to fix it? Or should it go and the functions changed so that they are throttled if the owner isn't in the sim?

I'm going to sit on it.

It needs to stay. It is under active development, and Andrew has stated that he will work with us to keep legitimate uses from staying broken. Several features he hinted at would make most of the problems with legit usage vanish, and several suggestions by various posters would cover [nearly?] all the rest.

I do not see this as an issue.

Toodle-oo!

I voted to keep it. I think the "doesn't apply if the owner is in the sim" is one of a number of logical and useful modifactions, but it doesn't solve the whole thing.

It would be a good one to implement quickly though.

Its rather a biased vote question... but I voted to keep it, because change to improve needs to be supported by residents, not whined into a stall by a vocal minority.

"Breaking existing content" is not a valid reason to not plug what LL has reported is a known problem and a reason the grid has been crashing. This computing "platform" is never going to grow into something better if every change has to support everything ever done in it.

That vocal minority is responsible for providing content for this world. This computing "platform" is never going to achieve sustainable growth if it doesn't begin offering users a more robust means of content creation, including better scripting tools.

Every major function in LSL has some sort of caveat, catch or drawback. If you pile on enough restrictions to prevent performance degradation or malicious intent, the language becomes so restricted that it begins to be useless for its primary purpose: creation of legitimate objects and scripts.

The ad hoc implementation of restrictions over the last couple of years has created a shoddy net. It does little to prevent malfeasance and nothing to prevent performance degradation. Instead, it gets tangled around the feet of this community's brightest members and prevents SL from being the environment it could be.

I voted to keep what we've got, and work on either adjusting the fence to become more "intellegent" or to provide a mechanism for exceptions to the fence (parcel / estate flag?)

Restricting llGiveInventory() for Obj to Obj transfer will break just as many legitimate objects as the current fence does.

So far three known broken scripts -- I know it sucks to suddenly have exisitng content break, and the forest rezzer (if it's really rezzing 65000 trees) would have to take nearly 30 minutes to stay under the fence. But if the current fence effectively lets the next grid wide attack be handled without the grid being taken offline, or reduces the time offline from 8hrs to under 2hrs -- then IMHO it's better to keep working with and adjusting the current fence.

Perhaps rather then messing with llGiveInventory(), the fence could simply not count objects that are rezzed on the owner's parcel(s)? It should still flag a warnning to the grid monkeys and to all the lindens currently on liason duty, so that a liason could drop in and check to see what's up?

One of the points I'm trying to make nobody has gotten yet. This change has broken existing content. There was no warning. Just bam! Broken. And later a post in the forums. Not all residents come here for one thing, and another is that this post was made in a section that most stay away from because they don't mess with scripts.

Actually Trep I think several of us have got the point - and whilst the 'fix' you're proposing would fix your content, and several of the other broken ones, hence why I'd like to see it SOON (Wednesday would be good if it needs serious downtime, rolling updates sooner if possible would be better, but changing the code will take time still) I think there are a number of other changes that could be sensibly made too, which is still why I voted for the other one. Thrashing out a "best possible" doesn't mean it has to be everything stays like this until it's finally decided, we can hope that Andrew will tweak, and continue to consult, to find a balance that is kind to legitimate content but makes grey goo attacks much harder.

A land box that lets us (not just at the estate level, at the parcel level) block external rez calls (default) or not is one I'd still like to see. You can bet most people would choose to block it or not change the default, and suddenly goo attacks find it much, much harder to find toe holds. It's the herd immunity approach, but depending on the internal structures of SL might be a lot harder to implement than the fence I guess.

Yes, this limits our creativity currently, but so does losing 16h in a weekend to two successive attacks. IMO the fence isn't a bad concept - but it does need tweaking (probably several sets of tweaking) to get right and at least Andrew is entering into the debate and it looks like he's taking the comments on board and will tweak away.

How many things got broken by the changes to permissions and teleporting - despite a short time when they said "oh we'll not do it then, until we aren't breaking things."? LL might aim to not break existing legitimate content, but it doesn't always succeed I'm afraid.