Jump to content

BIG DAY TODAY, passenger number 200,000 riding my vehicles.


You are about to reply to a thread that has been inactive for 2795 days.

Please take a moment to consider if this thread is worth bumping.

Recommended Posts

Yeah, but it's not completely trivial to keep track of the general direction the road is going in one sim in order to guess where it will go in the next one. It's certainly possible to guess correctly the vast majority of times, even on curves, aided by the fact that parcel boundaries come in 4m increments. Still, the smarter the script, the more memory it will use, so that's a trade-off; as a scripter, I might well choose less-smart behaviour if I could be sure they'd nearly never get stranded (perhaps if the temp-on-rez thing works).

Link to post
Share on other sites
  • Replies 263
  • Created
  • Last Reply

Top Posters In This Topic


AnnMarie Otoole wrote:

[...]

What, where is the "SLU thread"?  No one has contacted me regarding removal of them??

Has there been a change in temp-on-rez?  I actually spent quite some time with the Lindens trying to get the temp-on-rez altered.  The problem with it is the time is accumulative, if you disable and restore it continues from where it was so you only get about 60 seconds life.  I even filed a wish list to make the temp-on-rez resettable, or settable to a numeric time.  Have they done that? 

I also suggested a temp-on-no-script feature that would start a 1 minute self destruct timer if the object was on no-script land and this flag was set. 

 

I meant the thread to which you posted, on the SL Universe forums. (The snapshot I mentioned is in this post.)

The behavior that Maestro describes on comments to one of your jiras would be adequate to keep things temp-on-rez most of the time -- if it worked as reliably as he found in his testing. I don't know if it does, or if there are hidden gotchas. (The details of temp-on-rez have changed multiple times over the years, generally to contain griefer damage, but nothing all that recently as far as I know. I'm pretty sure you know all the changes that I know about, anyway.)

Link to post
Share on other sites

Nobody should have to be bothered to contact you regarding the prim litter you continue to spread from one corner of the mainland to another. It shouldn't be there in the first place, unless you have a method to ensure that it self-cleans, and avoids piling up all over the grid. You keep putting this on Linden Lab, saying it's their problem, because they aren't doing this or that or bending over backward to change the world to suit your annoying little fetish. A reasonable and responsible citizen of the Mainland would, under these circumstances, simply stop the program. What you're doing is virtually the same, ethically, as walking around your real world neighborhood, 24/7, taking a shi7 in your neighbors' yards, and arguing that until your shi7 figures out a way to clean itself up, it's your neighbors' responsibility to deal with it. 

The design intention for the temporary attribute was narrowly focused on expendables such as ordnance - bullets, missiles.. things that are intended to have single mission in a short life. It was not ever intended to accommodate what you are doing with these vehicles, and it is unreasonable and self-serving to expect that they would change the attribute's behavior to suit your personal interest.


AnnMarie Otoole wrote:


What, where is the "SLU thread"?  No one has contacted me regarding removal of them??

Has there been a change in temp-on-rez?  I actually spent quite some time with the Lindens trying to get the temp-on-rez altered.  The problem with it is the time is accumulative, if you disable and restore it continues from where it was so you only get about 60 seconds life.  I even filed a wish list to make the temp-on-rez resettable, or settable to a numeric time.  Have they done that? 

I also suggested a temp-on-no-script feature that would start a 1 minute self destruct timer if the object was on no-script land and this flag was set. 

 

Link to post
Share on other sites


Qie Niangao wrote:


AnnMarie Otoole wrote:

[...]

What, where is the "SLU thread"?  No one has contacted me regarding removal of them??

Has there been a change in temp-on-rez?  I actually spent quite some time with the Lindens trying to get the temp-on-rez altered.  The problem with it is the time is accumulative, if you disable and restore it continues from where it was so you only get about 60 seconds life.  I even filed a wish list to make the temp-on-rez resettable, or settable to a numeric time.  Have they done that? 

I also suggested a temp-on-no-script feature that would start a 1 minute self destruct timer if the object was on no-script land and this flag was set. 

 

I meant the thread to which you posted, on the SL Universe forums. (The snapshot I mentioned is in
.)

The behavior that Maestro describes on comments to 
 would be adequate to keep things temp-on-rez most of the time -- if it worked as reliably as he found in his testing. I don't know if it does, or if there are hidden gotchas. (The details of temp-on-rez has changed multiple times over the years, generally to contain griefer damage, but nothing all that recently as far as I know. I'm pretty sure you know all the changes that I know about, anyway.)

Well holy bananas Batman!

I just read https://jira.secondlife.com/browse/SVC-6725?

Ann starts a JIRA reporting a problem in February 2011!

Maestro suggests a fix.

Ann tries the fix and says it is not dependable.

Maestro does some more research and reports back that there is a setting that needs to be changed and that he found in his testing it is dependble but can't vouch 100%....he doesn't have time for more testing.

Ann, as it appears in that thread, instead of implementing the fix with Maestro's additional information, files a new JIRA, a feature request for a new function.

All this back in February 2011.

I see no indication that Ann ever tried to use Maestro's fix in her Vehicles.

It begins to look like Ann wants them to annoy everyone.

 

 

 

Link to post
Share on other sites

The problem isn't just sim crossings. The fundamental problem is that the type of navigation she is attempting to do simply cannot reliably be done with the resources at hand. There is nothing under the hood of LSL that will make this possible. The vehicles cannot stay on the road. They end up nudging into parcels set to disallow object entry, and often scripts as well. This renders them semi permanent fixtures across the Mainland landscape, until the parcel owners finally get around to visiting their property and removing the stalled vehicles. Crossing region boundaries is only part of the problem.

Anne just simply doesn't give a shi7. Her responses to criticisms in this thread and others indicates that she is acting with contempt for the community. That she believes it is appropriate to install rezzers for these vehicles on LDPW land, deliberately circumventing parcel restrictions designed to prevent it, and knowingly against the wishes of the Moles and Linden employees who have put so much effort and passion into creating an enjoyable Mainland experience confirms this.

This is not a technical problem. It isn't a programming problem. It is a behavior problem.


Gadget Portal wrote:

If the problem is sim crossings, just toggle temp for the crossing. Like I said before- crappy scripts.

 

Link to post
Share on other sites

In response to the suggestions that the Temp-On-Rez function may have changed since I tested it in 2011 or that Maestro Linden had supplied a solution I dug out the tester and tested again.  The only script function that is running in NO-SCRIPT land is the Temp-On-Rez timer that will destruct the object after about 1 minute.  The WIKI still says

"The garbage collection timer is cumulative and cannot be reset.  You can turn TEMP_ON_REZ off which stops the timer running but it will continue from where it left off if TEMP_ON_REZ is restored."

The theory is that you set TOR off for 1 second in every 60 so that there is a 59/60 chance that if it ends up on no-script land, the TOR is running, giving fairly good odds.  But with the tester off for 10 seconds in every minute, the longest time I could get it to run without destructing was 2 minutes.

Test it yourself, rez a cube and put this script in it.

 

integer life;integer temp_on_rez_Timer;default{    state_entry(){        llSetTimerEvent(0);        life = 0;        llSetPrimitiveParams([PRIM_TEMP_ON_REZ ,FALSE]);        llOwnerSay("Temp-On-Rez is OFF.  Click to start test.");    }        touch_start(integer t){        if(life == 0){            temp_on_rez_Timer = 0;            llSetTimerEvent(1);        }    }        on_rez(integer r){        llSetTimerEvent(0);        life = 0;        llSetPrimitiveParams([PRIM_TEMP_ON_REZ ,FALSE]);        llOwnerSay("Temp-On-Rez is OFF.  Click to start test.");    }    timer(){        life ++;        llOwnerSay((string)life);        temp_on_rez_Timer ++;        if(temp_on_rez_Timer == 1){            llOwnerSay("T-O-R is ON");            llSetPrimitiveParams([PRIM_TEMP_ON_REZ,TRUE]);        }        if(temp_on_rez_Timer == 51){            llSetPrimitiveParams([PRIM_TEMP_ON_REZ ,FALSE]);            llOwnerSay("T-O-R is OFF, being reset");        }        if(temp_on_rez_Timer == 60) temp_on_rez_Timer = 0;    }}

 

 

Link to post
Share on other sites


Gadget Portal wrote:

If the problem is sim crossings, just toggle temp for the crossing. Like I said before- crappy scripts.

No, need for cheap insults.

It takes more than a minute to cross sims, especially if they park so toggling TOR at the crossing still doesn't give you any more than a minute in the sim.  Sim crossings are NOT the problem.  I have virtually no sim crossing failures of unoccupied vehicles.

 

Link to post
Share on other sites

AnnMarie Otoole wrote:

Sim crossings are NOT the problem. 

I agree that sim crossing need not be a problem - I said that earlier. The real problem is that you are unable to create a system that actually works without littering the mainland all over the place. You keep blaming LL for the problems but, if it can't be done with the existing SL, it's not LL's fault - it's yours, because you continue to allow your system to litter the mainland, knowing that it just doesn't work properly. The only problem is you.

Nobody likes or wants your system. There may well have been a lot of people sitting on your vehicles - for a few seconds out of curiosity - but has anyone ever seen another person actually riding one? I.e. sitting on one and letting it give them a tour of the roads. In all the time I spent in a vehicle on the roads, I never saw anyone sitting on one of your vehicles. I did spend a fair amount of time trying to dodge them though, not always successfully, because they don't run smoothly along the roads, and there's no telling what they'll do next. They seem to have a habit of aiming at you. Nope. Nobody wants or even likes your vehicles system.

Link to post
Share on other sites


AnnMarie Otoole wrote:


Gadget Portal wrote:

If the problem is sim crossings, just toggle temp for the crossing. Like I said before- crappy scripts.

No, need for cheap insults.

It takes more than a minute to cross sims, especially if they park so toggling TOR at the crossing still doesn't give you any more than a minute in the sim.  Sim crossings are NOT the problem.  I have virtually no sim crossing failures of unoccupied vehicles.

 

Indeed, but here`s the problem as many stated befor, it is you and your junk

On the rare occsion i go to the mainland, every time i see a stupid ugly vehicle crashing around and when i checked out the railroad near hobo city, to my suprise there were a few of those ugly things going up and down the track that i just didn`t bother looking any further as your crap is littering every where

Link to post
Share on other sites


AnnMarie Otoole wrote:


Gadget Portal wrote:

If the problem is sim crossings, just toggle temp for the crossing. Like I said before- crappy scripts.

No, need for cheap insults.

It takes more than a minute to cross sims, especially if they park so toggling TOR at the crossing still doesn't give you any more than a minute in the sim.  Sim crossings are NOT the problem.  I have virtually no sim crossing failures of unoccupied vehicles.

 

Why is it taking them that long to cross a sim? i can cross a sim on my chopper in less than 20 seconds..

Link to post
Share on other sites


Drake1 Nightfire wrote:


AnnMarie Otoole wrote:


Gadget Portal wrote:

If the problem is sim crossings, just toggle temp for the crossing. Like I said before- crappy scripts.

No, need for cheap insults.

It takes more than a minute to cross sims, especially if they park so toggling TOR at the crossing still doesn't give you any more than a minute in the sim.  Sim crossings are NOT the problem.  I have virtually no sim crossing failures of unoccupied vehicles.

 

Why is it taking them that long to cross a sim? i can cross a sim on my chopper in less than 20 seconds..

Because of poor excuses

In Lionheart there is a vehicle dispenser from another creator and those cars cross within seconds, they`re still ugly but atleast don`t wander off the road or get stuck as they are instantly deleted if that happens or run out of road

Link to post
Share on other sites

I need to check the general discussion more often.

I had no idea this stuff was going on. I held back on doing any pathfinding stuff until I saw all the problems every1 else was having. Just thinking about it in the early beta stage, I could think of hundreds of potential problems, especially if the npcs weren't given a range from the rez point.

I went to a few different roads that had these vehicles, and I couldn't ride any of them for long, as the roads had fake constructions area which the huge vehicles could not navigate thru. Personally, I think the vehicles are way tooooooooo big. If I was going to do a project like this, I would start with little motor bikes, and maybe some rickshaws. Of course, anything that I would make would be animated too. My point is tho, that the vehicle should be half the size of the road. The larger you make an item, the harder times it's going to have with the Navmesh. I'd of not gone over only using half the width of the lane. Just do a test with a prim ramp, and see if you can get anything large to go up the ramp. It can't, because it runs into angles it can't fit thru. These are just my thoughts tho, I haven't played with pathfinding in months.

Link to post
Share on other sites

Yeah, but these vehicles are not pathfinding characters. They've been around since long before Pathfinding was even in beta, and to be honest, Pathfinding doesn't serve much purpose for what AnnMarie is trying to do here.*

Incidentally, I looked at that "T-O-R" script and played around with it. Adding a timestamp to the output shows that, in my tests, the object only ever got garbage collected if PRIM_TEMP_ON_REZ was turned on when one would expect garbage collection to occur: exactly 60 seconds after the object was set temp-on-rez. The script as written turns off temp nine seconds before gc, but then turns it on at the exact instant that gc should happen, but for whatever reason (accumulated float conversion error? timer event sloppiness? who knows) a little more than half the time that "exact instant" is a few microseconds before the 60s interval is up.

Switching the script to turn off temp-on-rez at 59 seconds and turn it back on at 61 (starting a new interval), the script has been running for me for several hours with no problem.

This, however, is in a static object, rezzed on a sim with tons of spare time, not sat-upon nor selected nor anything else that might well add noise to the garbage collector schedule, so I'm not asserting that this is necessarily a viable approach. I'm only saying that the specific script posted above doesn't demonstrate it to fail. (I'm also not seeing where Maestro's jira comments confirm the failure, but it's certainly possible that he came to that conclusion without updating the jira accordingly.)

_____

*It may be possible that the relatively new CHARACTER_STAY_WITHIN_PARCEL parameter to llGetStaticPath() might be useful in some variant of this, to find a valid route across a sim, but only given an a priori next-sim waypoint and a single Linden-owned parcel to navigate, neither of which exactly fits the problem as currently defined.

Link to post
Share on other sites


Qie Niangao wrote:

Yeah, but these vehicles are not pathfinding characters. They've been around since long before Pathfinding was even in beta, and to be honest, Pathfinding doesn't serve much purpose for what AnnMarie is trying to do here.*

Incidentally, I looked at that "T-O-R" script and played around with it. Adding a timestamp to the output shows that, in my tests, the object only ever got garbage collected if PRIM_TEMP_ON_REZ was turned on when one would expect garbage collection to occur: exactly 60 seconds after the object was set temp-on-rez. The script as written turns off temp nine seconds before gc, but then turns it on at the exact instant that gc should happen, but for whatever reason (accumulated float conversion error? timer event sloppiness? who knows) a little more than half the time that "exact instant" is a few microseconds before the 60s interval is up.

Switching the script to turn off temp-on-rez at 59 seconds and turn it back on at 61 (starting a new interval), the script has been running for me for several hours with no problem.

This, however, is in a static object, rezzed on a sim with tons of spare time, not sat-upon nor selected nor anything else that might well add noise to the garbage collector schedule, so
I'm not asserting that this is necessarily a viable approach
. I'm only saying that the specific script posted above doesn't demonstrate it to fail. (I'm also not seeing where Maestro's jira comments confirm the failure, but it's certainly possible that he came to that conclusion without updating the jira accordingly.)

_____

*It may be 
possible
that the relatively new CHARACTER_STAY_WITHIN_PARCEL parameter to llGetStaticPath()
might
be useful in some
variant
of this, to find a valid route across a sim, but only given an
a priori
next-sim waypoint and a single Linden-owned parcel to navigate, neither of which exactly fits the problem as currently defined.

Thank you Qui for testing.

Perhaps rezzing at a busy sandbox if you dont have another place to try would help 'stress test' the script.

One thing is how long does the script need to live with out failure?  And if it failed, what affect that would have on the vehicle in general.   After all, how long does it take for a vehicle to transit a continent before it would have auto deleted at the end of a trip.

At least this would help to reduce the problem.

 

Link to post
Share on other sites

Thanks a bunch, Qie.  I'll test it again.  It would be a major improvement if I can get it to work.  Prior testing (2011) had intervals varying from about 60 seconds up to 120 seconds with somewhat random times and no way to synchronize.  Perhaps (hopefully) subsequent releases have set the garbage collection to exactly 60 seconds. 

It would have to re-synchronize on sim crossings, I seem to remember that  you get a fresh garbage timer in each sim however that should not be difficult except some crossings take many seconds which may hurt timing..

 

EDIT

Doesn't work :(.

On a laggy sim I could not get it to survive 1 minute even with 5 seconds tolerance on synchronizing.

Link to post
Share on other sites


AnnMarie Otoole wrote:

Thanks a bunch, Qie.  I'll test it again.  It would be a major improvement if I can get it to work.  Prior testing (2011) had intervals varying from about 60 seconds up to 120 seconds with somewhat random times and no way to synchronize.  Perhaps (hopefully) subsequent releases have set the garbage collection to exactly 60 seconds. 

It would have to re-synchronize on sim crossings, I seem to remember that  you get a fresh garbage timer in each sim however that should not be difficult except some crossings take many seconds which may hurt timing..

you know what would be a major improvement? Removing you cars from the Linden roads. Why should i have to be pushed off the road by what amounts to a bot? Most people that have run into these things don't like them. I still want to know whose **** you suck to keep from being banned for improper usage of sim resourses. You know if one of us did this we would have been perma banned long ago. Tell Rod we say hi next time you are uner his desk.

Link to post
Share on other sites


Qie Niangao wrote:

Yeah, but these vehicles are not pathfinding characters. They've been around since long before Pathfinding was even in beta, and to be honest, Pathfinding doesn't serve much purpose for what AnnMarie is trying to do here.*


Of course, I'm not a coder, but I have played with pathfinding and it's purpose is exactly what AnnMarie is trying to do. The whole point of pathfinding is to find the path. Even tho I've helped develop animals that use only scripting, I would still prefer to use the pathfinding system, as it will work better for navigating than any script can do. Again, that is the point of pathfinding. These scripted touring systems have a problem finding the path and staying on it, because it has no navmesh to guide it. Again tho, I'm not a coder so I won't push the issue.

Link to post
Share on other sites


Medhue Simoni wrote:


Qie Niangao wrote:

Yeah, but these vehicles are not pathfinding characters. They've been around since long before Pathfinding was even in beta, and to be honest, Pathfinding doesn't serve much purpose for what AnnMarie is trying to do here.*


Of course, I'm not a coder, but I have played with pathfinding and it's purpose is exactly what AnnMarie is trying to do. The whole point of pathfinding is to find the path. Even tho I've helped develop animals that use only scripting, I would still prefer to use the pathfinding system, as it will work better for navigating than any script can do. Again, that is the point of pathfinding. These scripted touring systems have a problem finding the path and staying on it, because it has no navmesh to guide it. Again tho, I'm not a coder so I won't push the issue.

Which reminds me, I had read there were plans for the Moles to replace the current road beds with a Mesh Road.  I could be wrong about it being Mesh though.  See this thread:

http://community.secondlife.com/t5/Mainland/Strange-new-road-texture/m-p/1567813/highlight/true#M2840

I wonder if that will have any effect on the way her vehicles (are supposed to) stay on track.

Link to post
Share on other sites

Other than possible issues with the collision shape of the new road, if it's more than just a texture change, I doubt there'd be any considerable change in behavior. Most of the road pieces out there are just called "object", so I doubt her scripts are doing anything more than just trying to point in a general direction, with the name of the parcel under the wheels as the primary guidance reference.


Which reminds me, I had read there were plans for the Moles to replace the current road beds with a Mesh Road.  I could be wrong about it being Mesh though.  See this thread:

I wonder if that will have any effect on the way her vehicles (are supposed to) stay on track.

 

Link to post
Share on other sites


Medhue Simoni wrote:


Qie Niangao wrote:

Yeah, but these vehicles are not pathfinding characters. They've been around since long before Pathfinding was even in beta, and to be honest, Pathfinding doesn't serve much purpose for what AnnMarie is trying to do here.*


Of course, I'm not a coder, but I have played with pathfinding and it's purpose is exactly what AnnMarie is trying to do. The whole point of pathfinding is to find the path. Even tho I've helped develop animals that use only scripting, I would still prefer to use the pathfinding system, as it will work better for navigating than any script can do. Again, that is the point of pathfinding. These scripted touring systems have a problem finding the path and staying on it, because it has no navmesh to guide it. Again tho, I'm not a coder so I won't push the issue.

Although the phrase "finding a path" may describe both what Pathfinding does and what AnnMarie's vehicles do, they're two quite different problems. The vehicles are trying to discover directions to move so as to stay on Linden land, and generally on the long narrow parcels that correspond to roads (as opposed to random shapes and sizes of abandoned parcels, etc). I think they sometimes take advantage of objects rezzed on those parcels (rails, in particular) to help determine direction of motion, but they have to work on completely barren paths, too.

Or something like that. Anyway, the least of their challenges is to stay on the surface of the object/terrain navmesh, which Havok does for them because they're physical vehicles (as it would with Pathfinding).

The vehicles would behave differently -- and be much easier to program, with or without Pathfinding -- if they had at least one waypoint per sim, so the problem wasn't quite so under-constrained. Then, yeah, I could see some possible benefit to using that new "stay on a parcel" flavor of Pathfinding. (Roads are sometimes made up of different parcels within a sim, but given enough waypoints -- or enough other scripting to find intermediate waypoints -- it might yet be possible to find something useful for Pathfinding to do.) But again, that's solving the easy part of the problem; with no waypoints, Pathfinding characters can only wander using defined approach/avoidance parameters.

Link to post
Share on other sites

And given that those parameters have to be baked for each region with some regularity as the content geometry changes, it isn't something that can be relied on outside of a relatively controlled environment.


Qie Niangao wrote:

Pathfinding characters can only wander using defined approach/avoidance parameters.

 

Link to post
Share on other sites

Duh. I actually meant to post reporting about this: The simple 2 second non-temp interval I was testing eventually failed overnight. An earlier script is still running that stays non-temp for 5 seconds (and uses float constants to avoid LSL runtime typecasts, which are surprisingly compute-intensive). Probably, given enough lag, most any non-temp duration would eventually fail... and this is all kind of academic at this point because AnnMarie is testing a different approach suggested in the SLU thread.

One thing I wonder: What if the problem were changed such that the vehicles never even try to find their way back to the roads at all? If they find themselves off road, they log a failure message and self-destruct right then and there. This might unacceptably shorten trip lengths at first, until the scripts were tweaked to succeed more reliably at this more strict criterion, but I suspect this is actually a more tractable objective than trying to retrieve already stray vehicles without ever touching a problem parcel.

Link to post
Share on other sites
You are about to reply to a thread that has been inactive for 2795 days.

Please take a moment to consider if this thread is worth bumping.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...