Why I consider "path finding" useless in SL, and what do people use it for?

[161488303349]

i agree

pets are one of the few ways that non-maker people can participate in the wealth generating part of the economy

i had a fascinating chat with this person i random stumble into the other day

he is into breeding, training and racing SL thoroughbreds. is this quite large community into it. there are whole sim-size racetracks even

is amazing the kinds of money involved. .like this guy and other owners pay jockeys with superfast computers in the US to ride for them 1000s of lindens per race plus a cut of the prizepool. just like RL jockeys

this years Crown Cup which is the premier cup race paid 250,000L to the winner. there are heaps of other races as well where the race fee is like 3000L/4000L (some for only 100/200 for newbies) to enter your horse and 10s and 100s of 1000s in race prize money

on top of this he said the sidebetting is ferocious. like 100s of US$ that owners and race goers are betting with each other on the races

was extraordinary when i find out all this. we had a big chat about how to tune the viewer as well

 

  

[Mircea Lobo]

@ Innula Zenovka: Thanks for that link. It's nice to see even more being done on the side of graphics and rendering

@ Gadget Portal: Not really. I don't consider features bad because I don't use them. This one I consider useless because I couldn't think what anyone would use it for at all, but I see some people have plans. Only reason I'm upset about it is that "the package" came with removing OpenSim support, so I wanted to at least know what was so important to sacrifice it. Thought to be honest, like other people here stated, this might be better for OpenSim after all and set us free with our custom viewers.

[Ciaran Laval]

Pathfinding will evolve, the current issue for me is that we can't make NPC's with rigged skeletons. it makes progress in the area of NPC's extremely cumbersome and as such I'm not going to spend much time on it.

However as Pathfinding evolves and hurdles are overcome, people will embrace pathfinding. New gadgets and toys will arrive that become popular.

Builders will change the way they create builds to embrace pathfinding, they'll start doing it intuitively. However we're at an early stage still, but give it time.

[Perrie Juran]

---

Ciaran Laval wrote:

Pathfinding will evolve,

the current issue for me is that we can't make NPC's with rigged skeletons. it makes progress in the area of NPC's extremely cumbersome and as such I'm not going to spend much time on it.

However as Pathfinding evolves and hurdles are overcome,

people will embrace pathfinding. New gadgets and toys will arrive that become popular.

Builders will change the way they create builds to embrace pathfinding, they'll start doing it intuitively. However we're at an early stage still, but give it time.

---

It has a long friggin way to go.

A very long friggin way.

[solstyse]

---

Gadget Portal wrote:

What I get from this forum thread is "I can't think of a way to use Pathfinding and I don't like products that do, so it's a useless feature! MY feature requests are FAR more important!"

---

I have to disagree slightly. What I get from this thread is "Once again, LL rolled out a feature before it was perfected and failed to convince me that it's worth more than things I know people have been asking for."

Pathfinding has a long way to go. Most npc's that I see are pitiful. And land featuring boxes that move is a terrible way to sell something to a customer. It may please a focus group, since they are looking for a sort of prototype. But by the time something reaches the end consumer (us) it should be a finished, polished product. Often, what LL gives us is not. Often, once 90 percent of the bugs we first see are worked out, 90 percent of us finally embrace their new ideas.

Good npc's that move well are basically what I think it would take for pathfinding to gain widespread acceptance. The people wwith animals that see it as a way to lower their script count really are kind of a niche. For now, most sims I see would probably be best with pathfinding disabled.

[Innula Zenovka]

---

solstyse wrote:

For now, most sims I see would probably be best with pathfinding disabled.

---

Why?    What criteria are you using?  

I'm not saying you're mistaken; it's just I certainly wouldn't know how to reach such a diagnosis.   What figures should I be looking at, and what should lead me to say "this is the problem and the way to improve things is probably to turn pathfinding off, rather than to do something else"?

ETA -- the advantage of pathfinding,  from the scripting point of view, isn't that it lowers the script count (though it probably does) but that it performs some tasks considerably economically than the methods we've had to use before.     The sim is having to do a lot less work to acheive the same result, and it's also got dedicated resources for that task, so it's not cutting into the script time available for other tasks.   That's the bonus rather than the amount of memory you're using or not.

[Pussycat Catnap]

---

16 wrote:

---

Jadeclaw Denfu wrote:

That's one purpose of pathfinding:

http://maps.secondlife.com/secondlife/Jelloab/125/82/52

Collision avoidance. Watch the cubes avoiding the pillars, each other and you.

Yes, try to walk into their way, the cubes will go around you or turn away.

(If the cubes don't move when you visit, click them once to get them moving.)

 

---

^^ that ^^

one thing that brings sims to their knees the most is collisions. the next most is running detectors to avoid collisons

if we can have an easy way to not have to do either then is better in the long run

---

If pathfinding had been around when scion chickens launched - nobody would ever have developed a negative bias against breedables (I used to AR the things... now I have a lot of them)...

- They used to be severly laggy, due to the original versions having to deal with all kinds of collision issues.

The general solution now is that most of them are phantom, and walk on flat plains; making them awkward on hills or in houses with rises or depressions.

Pathfinding would have been ideal - and will probably feature into breedables once everything's ready / the next generation of breedables comes along.

 

And yeah on Havok. Any old Open Sim could I imagine go out and buy its own Havok license. Its not LL's job to pay their bills.

 

EDIT:

---

JobsFox wrote:

Pathfinding is FAR from useless. In fact it solves some HUGE problems for product developers and estate managers in SL.

Snuffles is the first breedable that uses the pathfinding system. They generate hardly any lag because they do not use old collision based movement systems. 

http://www.goldtokens.net/index.php?title=Snuffles

---

Guess I need to read farther before I post. I will have to go check these out.

 Can't afford to maintain another breedable at the moment, but I might play around with them if interesting.

 

[161488303349]

a barge or gondala system could work. navigate round small obstacles in the canals. things like that maybe

whitewater rafting/kayaking maybe also. dunno the exact details how to implement. but say the kayak follow a path and avoids rocks to emulate following the flow of the water

then can use the arrow keys to apply extra/less power/direction to the kayak. could even make rapids with elevator prims. whirlpool even that suck you down so you have to apply more power to escape

am just thinking about whatifs and whatelses

[Qie Niangao]

---

16 wrote:

a barge or gondala system could work. navigate round small obstacles in the canals. things like that maybe

whitewater rafting/kayaking maybe also. dunno the exact details how to implement. but say the kayak follow a path and avoids rocks to emulate following the flow of the water

then can use the arrow keys to apply extra/less power/direction to the kayak. could even make rapids with elevator prims. whirlpool even that suck you down so you have to apply more power to escape

am just thinking about whatifs and whatelses

---

I'm dubious about using pathfinding to avoid static obstacles on a pre-defined course; pathfinding may be efficient, but it seems wrong to force it to solve the exact same problem over and over again. The potential win, I think, is for avoiding dynamic obstacles such as avatars and other vehicles, and (at least theoretically) for finding a navigable path on a new course. This didn't seem to work at all reliably when I tried early on, but maybe things are improved enough now to warrant another try. Or maybe I was just doing it wrong.

[Phil Deakins]

---

Mircea Lobo wrote:

@

Gadget Portal

: Not really. I don't consider features bad because I don't use them. This one I consider useless because I couldn't think what anyone would use it for at all, but I see some people have plans. Only reason I'm upset about it is that "the package" came with removing OpenSim support, so I wanted to at least know what was so important to sacrifice it. Thought to be honest, like other people here stated, this might be better for OpenSim after all and set us free with our custom viewers.

---

I'm curious as to what sort of support LL gave to OpenSim, other than moral support. I thought that OS was only a copy of SL that was written by the OS programmers (reverse engineered) and that LL has never released server code. They released viewer code but I don't think they released server code. Did that change?

In my view, once the OS developers had repreoduced SL, they should have been off on their own, developing it quite seperately from SL - making a genuine alternative.

[Phil Deakins]

Question: The pathfinder tools were only going to be available to selected people - presumably those who actually build and apply for them. Is that still the case, or does everyone now have access to them?

[Mircea Lobo]

@ Phil Deakins: The only "support" they gave to OpenSim was allowing the --loginuri parameter in their viewer, and usually no one ever asked them for more. But now they removed that as well... though thank god the SL viewer is open-source so there will always be viewers with custom grid support. Still, it's sad. Hence my colder attitude about path finding support, considering it likely wasn't a very crucial feature either (and they could have found ways to implement it without making up crazy and silly licensing claims).

[Phil Deakins]

So LL didn't support OpenSim, other than years ago, when they said they supported the effort - moral support.

LL releases viewer code but I'm not up to date with how they deal with third party viewers. Does LL have a list of TPVs and only those viewers can be used with SL? If that's the case, has LL removed the OS viewer from the list? If LL is now preventing the OS viewer, there must be a reason, and it can't be merely that pathfinder exists, because other viewers are not prevented.

Either way, OS continues - that's OpenSim itself - not a viewer. It never needed LL's cooperation so it seems to me that nothing has changed.

[161488303349]

i was thinking more about the water.  at the moment (without navmesh) SL water is either flat or we make a tube ride

RL water flows and changes. erodes rocks, brings down cliff faces, creates sandbars. creates and erodes obstacles. then there are high water and low water seasons and so on

i was thinking how could that be done with navmesh. and how could we put a boat/kayak/raft/gondola/barge on the water. and how/what control would we give the avatar on it, given that the boat is going to go with the flow which would/should change over time

+

can understand tho the animal on animal thing. hunter vs prey. swarms, etc. tho like Ciaran mention until we get rigged and bones then is not very realistic

was just thinking of what else can/could be done now at the moment other than maze and sniffers

 

[Qie Niangao]

---

Phil Deakins wrote:

Question: The pathfinder tools were only going to be available to selected people - presumably those who actually build and apply for them. Is that still the case, or does everyone now have access to them?

---

Not sure whether it was really ever the plan that pathfinding was to be among those tools that would only be available to folks who registered for them. There was a time, of course, when pathfinding was in beta, but I can't remember what the rules were then. (If participation required an NDA, I wasn't involved, that's for sure.)

What I thought was going to be limited to registered users was the "Experience Permissions" stuff, which seems to be quite slow to emerge. There was a recent faint glimmer of hope, but nothing definite. They may have gotten spooked when in-world Lindens were force-teleported to The Corn Field sims as llTeleportAgent() was first deployed with very braindamaged permissions. (Now, of course, that function has been nerfed to nearly total uselessness.)

 

---

@16: Yeah, it's good to brainstorm more interesting applications of the technology. Somehow, I always end up frustrated, wanting to get "underneath" the script API. Water is actually a good example: it's frustrating that the navmesh simply stops at llWater()'s edge. (I know you were talking about synthetic water, but the llWater() problem defeats some ideas I might have had.) Another notion, before any API was defined, was to limit path navigation to a particular parcel or parcel owner (Governor Linden), but that's pretty far from what's possible now.

[Solar Legion]

---

Mircea Lobo wrote:

@

Phil Deakins

: The only "support" they gave to OpenSim was allowing the --loginuri parameter in their viewer, and usually no one ever asked them for more. But now they removed that as well... though thank god the SL viewer is open-source so there will always be viewers with custom grid support. Still, it's sad. Hence my colder attitude about path finding support, considering it likely wasn't a very crucial feature either (and they could have found ways to implement it without making up crazy and silly licensing claims).

---

Linden Lab did not "make up" any sort of "crazy and silly" claim. The makers of Havok are the ones that reminded Linden Lab of their agreement/Terms of Use for the Havok engine.

Sorry but ... that is something I could not remain silent on any longer. Make sure you have your facts straight before launching into a diatribe over a loss of "support" that was never really there to begin with.

[Innula Zenovka]

---

Phil Deakins wrote:

LL releases viewer code but I'm not up to date with how they deal with third party viewers. Does LL have a list of TPVs and only those viewers can be used with SL? If that's the case, has LL removed the OS viewer from the list? If LL is now preventing the OS viewer, there must be a reason, and it can't be merely that pathfinder exists, because other viewers are not prevented.

---

Separately from releasing the viewer code, LL allows approved TPVs indirect access to the closed source Havok technology; see the wiki official article on the Havok Viewer Sublicence.    The actual Sublicence agreement that devs are required to sign actually goes further than that, saying that the "Sublicensee must require the Third Party Viewer to connect only to servers owned or operated by the Company" (i.e. LL).

Separately, but in an apparently related measure, LL removed support for the -loginURI parameter in the official viewer, which means that you can no longer use the LL viewer to connect to non-LL grids and that TPVs have to put the parameter back in if they want their viewer to connect to other grids.

Some TPVs (e.g. Firestorm) are producing two versions, one -- with Pathfinding support and without the -loginURI parameter -- for the main grid and one, without Pathfinding support, that will connect to other grids.   Other viewers have decided that they can't maintain two separate branches of the viewer, so have dropped support for other grids.

In effect, all this means is that if I want to log in to Open Sim I have to use something like Kokua (the V3 successor to Imprudence) or  the Open Sim edition of Firestorm, and when I get to one of these grids I  won't have access to the viewer functions to manipulate the navmesh that I would need if I were setting stuff to being walkable or not on the main grid.   Since I don't think Open Sim has pathfinding anyway, that doesn't seem any great drawback -- particularly, I would have thought, to someone like Mircea who regards pathfinding as a waste of time anyway.

TL;dr.  If you want to log into Open Sim you have to use a TPV, and possibly a special edition of one.   Nothing to stop OS devs compiling their own viewer primarily  for use by people who want to access third party grids, of course.

[Ciaran Laval]

---

Qie Niangao wrote:

I'm dubious about using pathfinding to avoid static obstacles on a pre-defined course; pathfinding may be efficient, but it seems wrong to force it to solve the exact same problem over and over again. The potential win, I think, is for avoiding

dynamic

obstacles such as avatars and other vehicles, and (at least theoretically) for finding a navigable path on a new course. This didn't seem to work at all reliably when I tried early on, but maybe things are improved enough now to warrant another try. Or maybe I was just doing it wrong.

---

I agree largely with you that pathfinding to avoid static obstacles repeatedly isn't ideal, but from a merchant perspective, when you don't know how people are going to use your item, it probably has some end user pleasing advantage, but yes, pathfinding should be about obstacles whose path is more difficult to predict.

[Ciaran Laval]

---

Phil Deakins wrote:

Question: The pathfinder tools were only going to be available to selected people - presumably those who actually build and apply for them. Is that still the case, or does everyone now have access to them?

---

 Pathfinding went into beta and you could apply to have it installed on your sim, there was a special pahtfinding viewer. Obviously as you had to apply to have it installed on your own sim, it was restricted, but it wasn't only intended to be available to certain people.

I think you may be getting confused with Linden Realms and the tools they have been using there, they did use pathfinding on Linden Realms but they also had advanced experience tools such as teleport agent and temporary attachments, these were the tools that were being mooted as requiring some sort of creator program to use.

[Mircea Lobo]

@ Solar Legion: When OpenSim support was dropped, Linden specified that the reason they did that was because "The owners of Havoc would no longer allow them to connect to custom grids". Sorry, but am I the only one who finds that the one of the silliest claims in SL history?

Havoc does not work at Linden Lab. They make a physics engine library for games, which they then sell to game companies that wish to use them. They normally have no authority over the development of SL or what the SL code does. I'd be amazed if anyone over at Havoc even knows what Second Life is in detail, what a simulator is, what --loginuri does, etc. Sure, the owner of the Havoc library get to say "You can include my work with this project or not", but I've never heard of an external person having the authority to give orders to a project they don't work at or know anything about, even in order to allow usage of their library.

I still tend to believe this was just an excuse to cut OpenSim because it was ruining SL grid business. But in case I am wrong and what LL says was true behind closed doors, LL has was a very naive company to give authotiry to other corporations in the development of their project. Personally at least, I never heard of such a thing before. But as far as I'd understand, this makes Havoc part of the SL development team in that case.

[leliel Mirihi]

---

Innula Zenovka wrote:

---

Phil Deakins wrote:

LL releases viewer code but I'm not up to date with how they deal with third party viewers. Does LL have a list of TPVs and only those viewers can be used with SL? If that's the case, has LL removed the OS viewer from the list? If LL is now preventing the OS viewer, there must be a reason, and it can't be merely that pathfinder exists, because other viewers are not prevented.

---

Separately from releasing the viewer code, LL allows approved TPVs indirect access to the closed source Havok technology; see the wiki official article on the

Havok Viewer Sublicence

.    The

actual Sublicence agreement

that devs are required to sign actually goes further than that, saying that the "Sublicensee must require the Third Party Viewer to connect only to servers owned or operated by the Company" (i.e. LL).

Separately, but in an apparently related measure, LL removed support for the -loginURI parameter in the official viewer, which means that you can no longer use the LL viewer to connect to non-LL grids and that TPVs have to put the parameter back in if they want their viewer to connect to other grids.

Some TPVs (e.g. Firestorm) are producing two versions, one -- with Pathfinding support and without the -loginURI parameter -- for the main grid and one, without Pathfinding support, that will connect to other grids.   Other viewers have decided that they can't maintain two separate branches of the viewer, so have dropped support for other grids.

In effect, all this means is that if I want to log in to Open Sim I have to use something like Kokua (the V3 successor to Imprudence) or  the Open Sim edition of Firestorm, and when I get to one of these grids I  won't have access to the viewer functions to manipulate the navmesh that I would need if I were setting stuff to being walkable or not on the main grid.   Since I don't think Open Sim has pathfinding anyway, that doesn't seem any great drawback -- particularly, I would have thought, to someone like Mircea who regards pathfinding as a waste of time anyway.

TL;dr.  If you want to log into Open Sim you have to use a TPV, and possibly a special edition of one.   Nothing to stop OS devs compiling their own viewer primarily  for use by people who want to access third party grids, of course.

---

I'm shure some people will cry foul at this but in the long run I think it's for the best. OpenSim was never going to 'go' anywhere as long as they were just following in SL's foot steps. I hope this will motivate them to spread their wings as it were and innovate, instead of just being a me too clone that's months behind. Not to say they haven't innovated in some ways, but maintaining compatibility with SL limits what they can do quite a bit.

[Phil Deakins]

Thank you for that full explanation, Innula. So the idea that LL cut the OS viewer off is just a red herring because they didn't treat it any differently to the other TPVs. The OS people could make 2 versions, like the Firestorm people do, and all would be well.

[Mircea Lobo]

After a while, I've actually come to agree with this. It might be best for OpenSim in the end, since now they can probably make their own client and add viewer-side features that way (I don't know if that direction will be decided however). Still upset with LL simply cutting us off this way though, and this is a big loss at the same time.

[Phil Deakins]

Yes, I'm mistaking the pathfinder stuff with the tools that make Realms work. I assume that the pathfinder stuff is in LSL, whereas the Realms tools are actually tools. Is that right?

[Czari Zenovka]

---

Mircea Lobo wrote:

After a while, I've actually come to agree with this. It might be best for OpenSim in the end, since now they can probably make their own client and add viewer-side features that way (I don't know if that direction will be decided however). Still upset with LL simply cutting us off this way though, and this is a big loss at the same time.

---

What viewer/s can now be used for OpenSim and do you recommend any specific one?  I have visited OSGrid a few times since a group of my friends left SL and set up their new online home there.  I accessed it with Phoenix.  Haven't tried that in awhile so it may no longer work.

Edited for spelling