User:Andrew Linden/Office Hours/2008 10 07

Transcript of Andrew Linden's office hours:

[11:05]	Kitto Flora:	Hello
[11:05]	Andrew Linden:	Hello. I'm only here on time because Simon Linden reminded me about the hour!!
[11:05]	Orion Shamroy:	hehe ;) Good morning :)
[11:05]	Andrew Linden:	We were in the middle of a code review just a few minutes ago.
[11:05]	Kitto Flora:	And Simon fogot?
[11:05]	Andrew Linden:	No, he actually remembered. I had forgotten
[11:06]	Kitto Flora:	Did the camera control fix get into the latest RC?
[11:06]	Andrew Linden:	No Kitto, that fix is a server-side thing
[11:06]	Andrew Linden:	so it won't be available until 1.25-Server
[11:06]	Kitto Flora:	Oh
[11:06]	Rex Cronon:	hello everybody
[11:06]	Andrew Linden:	Hi Rex
[11:06]	Kitto Flora:	Maybe I didnt go places where its a problem, but it seems better
[11:06]	Orion Shamroy:	Is that the thing that keeps making my cam go bonkers when I'm in a tight space?
[11:06]	Kitto Flora:	Hello
[11:07]	Andrew Linden:	Rex, you seem to have settled on the Iron Man avatar.
[11:07]	Peter Bamaisin:	Hey.
[11:07]	Andrew Linden:	You think maybe it is going to stick?
[11:07]	Kitto Flora:	Orion: Yes
[11:07]	Rex Cronon:	i like it:)
[11:07]	Orion Shamroy gets on his knees and praises Andrew... "Thank you thank you!" <<throws away his bottle of dramamine>> :)
[11:07]	Kooky Jetaime:	I'm shocked they havn't put in more updates in 1.20 than just the three lines of code....... this is soo..not typical Linden..
[11:08]	Andrew Linden:	Unfortunately the camera oscillation bug was fixed months ago... and is still not yet deployed.
[11:08]	Rex Cronon:	sadly i can't wear this in rl:)
[11:08]	Kooky Jetaime:	And people wonder why users are frustrated with LL
[11:09]	Kitto Flora:	Someone.. .wonders?
[11:09]	Kooky Jetaime:	I was being nice :)
[11:09]	Andrew Linden:	Really Kooky? I didn't think there was much of a mystery there ;-)
[11:09]	Orion Shamroy:	Well, I'm sure it happens in large projects like the SL server... There's alot of components and stuff to be done. :)
[11:09]	Kitto Flora:	Is there news?
[11:09]	Kooky Jetaime:	But it makes you wonder how many other fixes have gotten lost in the beuracracy
[11:09]	Andrew Linden:	There are a few people who "understand".
[11:10]	Orion Shamroy hides behind his "C# for VB Dummies" book...
[11:10]	Rex Cronon:	this camera bobbing, reminds me of how "lara croft" camera used 2 work:)
[11:10]	Kooky Jetaime:	hahah
[11:10]	Kitto Flora:	8 months ago I reported on Jira that the beach-ball task was broken on OI. Yesterday some Linden responded. *+ months to review a Jira bug!
[11:10]	Kooky Jetaime:	Careful Rex, LL will hear that then call it a "feature"
[11:10]	Kooky Jetaime:	They'll call it "Tomb Raider" Mode
[11:11]	Andrew Linden:	Oh right! I remember an oscillation bug in one of the original "Tomb Raider" games, on the PlayStation
[11:11]	Rex Cronon:	that is way i only played that game 4 only 15 minutes
[11:11]	Andrew Linden:	I don't have much news...
[11:12]	Kooky Jetaime:	I'm waiting for the clothing glitch to get fixed
[11:12]	Andrew Linden:	looks like 1.25-Server will be delayed some more by some emergency fixes that went out this weekend and yesterday, and ate up a bunch of QA resources.
[11:13]	Andrew Linden:	I was suggesting to the #deploy channel this morning that they just branch for 1.25-Server now and deliver the changes that have piled up for months. They weren't too keen on that... mostly to conserve QA resources
[11:14]	Andrew Linden:	the QA team is already busy busy, with queues of fixes that need to "tested for fixed"
[11:14]	Andrew Linden:	or tested again for mergeback (regression)
[11:14]	Kitto Flora:	sounds like QA is broken.
[11:14]	Andrew Linden:	well, not real regession in that final step
[11:14]	Kooky Jetaime:	Or understaffed/sized
[11:15]	Andrew Linden:	QA is just overloaded
[11:15]	Kitto Flora:	Ineffecive too?
[11:15]	Kitto Flora:	Yay Simon
[11:15]	Simon Linden:	Hi everyone
[11:16]	Andrew Linden:	Relatively effective. Some may be able to remember worse QA problems way back in 1.5 - 1.9 era
[11:16]	Peter Bamaisin:	Hi
[11:16]	Andrew Linden:	Hello Peter, take a seat.
[11:16]	Andrew Linden:	That is, if you are here for "Andrew Linden's office hours"
[11:16]	Rex Cronon:	hi
[11:17]	Kooky Jetaime:	Oh ya, I remeber that era
[11:17]	Kooky Jetaime:	Weekly updates too
[11:18]	Kitto Flora:	I havea bug to look at - its not h4 related ... though it may have been introduced around the start of H4. Its year+ old
[11:18]	Andrew Linden:	What bug is it?
[11:18]	Kitto Flora:	SVC-998
[11:18]	Kitto Flora:	click the cubes for a Jira Link
[11:19]	Kitto Flora:	Its about what the axis arrows show when editing a linked set in Local Ruler Mode
[11:19]	Andrew Linden:	hrm... I recall reading that one a while ago
[11:19]	Kitto Flora:	Its still there
[11:20]	Andrew Linden:	I think that is a viewer bug
[11:20]	Kitto Flora:	One might expect that axis arrows would show the orientation of the ROOT prim.
[11:20]	Kooky Jetaime:	Kitto, does that have anything to do with the one where the 6-way arrows and the viewer center-top location display mis-match the location in the editing tool?
[11:20]	Kitto Flora:	BUT they show ... well it varies on the situation. Its probably nere root.
[11:21]	Kitto Flora:	Kooky - it may well be related
[11:21]	Kitto Flora:	I'd expect the arrows to be cenetred on the root prim - but again - they are not
[11:21]	Kooky Jetaime:	I remember having an issue when moving an object and it was mis-matching
[11:21]	Kitto Flora:	Yep - its borked
[11:21]	Kitto Flora:	for .. a year or so now
[11:21]	Kooky Jetaime:	Why would they be centered on the root and not the object?
[11:22]	Andrew Linden:	No, selection arrows are located in some average center of the linked set
[11:22]	Kitto Flora:	Well - the linked object itself has no 'rotation' - only the prims in it have a rotation
[11:22]	Kitto Flora:	Location - yeah - Average center would be good
[11:22]	Kooky Jetaime:	uhm.
[11:22]	Kitto Flora:	Its the oreintation thats the problem
[11:23]	Andrew Linden:	well, one might expect the arrows to jump to the root upon link
[11:23]	Kooky Jetaime:	if you rotate a linked object, the entire object rotates around that center point, not each prim around its own center, so how does an object not have rotation?
[11:23]	Andrew Linden:	hrm... right kooky
[11:23]	Andrew Linden:	anyway, the orientation is the bug, and it is incorrect
[11:23]	Kitto Flora:	The displayed local orientation is one thing when you first link it
[11:24]	Kitto Flora:	If you take and re-rez it - the orientation changes
[11:24]	Kitto Flora:	And if you rez it somewhere else it changes again
[11:24]	Andrew Linden:	there is a possibility that this is somewhat related to the server... perhaps it can be fixed server-side by forcing an update... although I would expect those updates were already sent upon linkage
[11:24]	Kitto Flora:	For me its now aligned with link 4
[11:25]	Kitto Flora:	But last time I looked at this it was link 2
[11:25]	Kitto Flora:	Andrew: Any idea where the info for the orientation for the arrows comes from? Server or client?
[11:25]	Andrew Linden:	alright kitto, I'll make sure the bug has been imported to the internal list, and I'll try to talk to a dev who knows more about how it.
[11:26]	Andrew Linden:	er... how it is supposed to work
[11:26]	Kitto Flora:	Well - I would expect local mode to show the orientation of the root prim
[11:27]	Andrew Linden:	On a related but distinct note... perhaps the client UI should be toggleable to keep the selection point on the root, or the to-be-root object
[11:27]	Andrew Linden:	however... at the moment the to-be-root object is the *last* selection instead of the first
[11:27]	Kitto Flora:	Last is good
[11:27]	Kitto Flora:	Means its easy to change the root object
[11:27]	Kitto Flora:	Er
[11:27]	Kitto Flora:	Root prim
[11:28]	Andrew Linden:	Right, you can change it without having to reselect everything in the exact right order
[11:28]	Kitto Flora:	Eggzkly
[11:29]	Andrew Linden:	Anybody here ever file a DMCA takedown request? Just curious
[11:29]	Orion Shamroy:	0-o Not I?
[11:30]	Kooky Jetaime:	Nope
[11:30]	Andrew Linden:	I was talking with another developer about how our DMCA takedown toolset should work. I was curious if anyone had had experience with it. Oh well, no worries.
[11:30]	Kitto Flora:	Yo need a RL corporation goon.
[11:30]	Rex Cronon:	u should ask that question on a builders group
[11:31]	Kitto Flora:	Who else would care about DCMA?
[11:31]	Andrew Linden:	Right.
[11:31]	Kooky Jetaime:	Nods
[11:31]	Rex Cronon:	or groups that make textures
[11:32]	Rex Cronon:	oh, i know who u should ask:)
[11:32]	Rex Cronon:	that guy with sex bed, that had to sue in rl:)
[11:32]	Orion Shamroy blushes for knowing this... "Stroker Serpentine"
[11:33]	Rex Cronon:	right
[11:33]	Andrew Linden:	Maybe I'll bother Stroker if I find the time.
[11:34]	Andrew Linden:	If anyone has some questions or an issue, feel free to mention it. I didn't bring much of an agenda.
[11:34]	Rex Cronon:	so, any new limitations/fixes with the new viewer?
[11:35]	Kitto Flora wonders if Andrew looked at that bouncing vehicle from Leendert
[11:35]	Andrew Linden:	Not yet Kitto.
[11:35]	Rex Cronon:	maybe we finally have mega-prim creation:)
[11:35]	Andrew Linden:	the brand new viewer had some security fix in it.
[11:35]	Andrew Linden:	I didn't work on that project.
[11:35]	Orion Shamroy:	megaprims ++ :)
[11:36]	Andrew Linden:	oy
[11:36]	Andrew Linden:	Alas, no progress on my megaprim liberation project :-(
[11:38]	Andrew Linden:	I now realize that the Havok4 project was a long lull between the normal more frantic environment where the important projects show up way too fast.
[11:38]	Rex Cronon:	oh, and on collision, when u use lldetected pos, could we get the position in sim coordinates , of where the collision took place?
[11:38]	Rex Cronon:	right now it returns the current position
[11:38]	Andrew Linden:	Actually Rex that is somewhat related to a problem Simon had been working on a few weeks ago
[11:39]	Andrew Linden:	Simon Linden had been working on resurrecting collision particle sprites
[11:39]	Simon Linden:	Ah, that was a few days wasted :(
[11:39]	Kitto Flora:	One wonders what the 'important projects' are. Mysterious securty patches? Its sure not the bugs we all keep asking about.
[11:39]	Andrew Linden:	however there is no "offset" property for object particle system creation.
[11:39]	Rex Cronon:	particles colliding with objects?
[11:39]	Andrew Linden:	no Rex, we used to have collision puffs of smoke
[11:40]	Simon Linden:	There used to be 'dust' and such created when some things collided hard
[11:40]	Kitto Flora:	Did they break?
[11:40]	Andrew Linden:	so if an object hit the ground it produced some particle sprites at the collision point
[11:40]	Kooky Jetaime:	Andrew, quick question before we move off fo megaprims, Now we're on H4, is there any performance disadvantage of using already existing megaprims?
[11:40]	Kitto Flora:	I cant recall seeing any recently
[11:40]	Andrew Linden:	yeah, they broke more than a year ago
[11:40]	Rex Cronon:	i think u can replace the collision particles
[11:40]	Simon Linden:	It went away over a year ago
[11:40]	Andrew Linden:	when the "legacy particle system" was ripped out
[11:41]	Andrew Linden:	anyway, world position was needed for the particles... not local position
[11:41]	Andrew Linden:	and Rex is asking for world location info in collision events
[11:41]	Simon Linden:	The other problems is that particle systems are now always tied to an object
[11:42]	Kitto Flora:	Oh - I do see that *apparenty* the bug with partciles 'not aligning with the source' in dynamic objects has been fixed. But the wrong way :)
[11:42]	Rex Cronon:	i think that is not that unusual to know where in the world did 2 objects colliede.
[11:43]	Orion Shamroy:	Kookey, I've noticed better performance / less weirdness with the new set of megaprims that came out. The ones that dont use all those weird path cuts and such.
[11:43]	Andrew Linden:	Rex, what would the interface look like?
[11:43]	Andrew Linden:	How would you inform the script you were interested in WorldPosition instead of LocalPosition?
[11:43]	Kooky Jetaime:	New set? I didn't know the old set had strange cuts unless you gave it to them
[11:44]	Rex Cronon:	when u call from inside the collision_end, lldetectedpos, u should get the position in world coordinates where the collision took place
[11:44]	Orion Shamroy:	New as in the ones that werent created by Gene Replacement.
[11:45]	Andrew Linden:	But Rex, we can't change the way it works now (since untold amounts of content would then break) so we would have to add new functionality
[11:45]	Rex Cronon:	i guess u migth need a new function lldetectecCollisionPos
[11:45]	Arawn Spitteler:	llDetectedPos is not giving Global Co-ords?
[11:46]	Rex Cronon:	arawn, is giving them, but that is the pos of object right now, not where collision took place
[11:46]	Kitto Flora:	Its returns the position of the prim in which is the script
[11:46]	Andrew Linden:	ah right... hrm
[11:46]	Kitto Flora:	can one get info on the other object - like where it is?
[11:47]	Andrew Linden:	the actual collision location is probably not even saved or stored anywhere in memory where the LSL callback could get it...
[11:47]	Arawn Spitteler:	It should return the position of that object, that collided, but we alson need a feature, to detect characteristics of the moment of event, rahter than at the time the event was called
[11:47]	Rex Cronon:	where it is at time of asking, but not were it was at time of collision
[11:47]	Andrew Linden:	we'd have to add some data to some structure
[11:48]	Andrew Linden:	and then those scripters who wanted to get at the actual world collision location would have to use some different llDetected* call
[11:48]	Rex Cronon:	lets say u want a boar for darts, u need to know where the dart was at the time of the collision
[11:48]	Rex Cronon:	a board of darts*
[11:49]	Kitto Flora:	It occurs to me that if you can find the position of the colliding object, one might estimate the contat point
[11:49]	Arawn Spitteler:	What's the feasibility of Events containing an object of data
[11:49]	Rex Cronon:	if both r physical, is kind of hard
[11:49]	Andrew Linden:	well Rex... the current system would work fine if the collision event were handled in a timely manner
[11:50]	Andrew Linden:	that is, the dart would not have moved very far from where it actually collided... if the collision event were hanlded in a timely manner
[11:50]	Arawn Spitteler:	At CG's, someone was talking about arrows that would bounce 5 meters up, before the event was handled
[11:50]	Andrew Linden:	yes, that was a bug that was related to SVC-2511
[11:50]	Rex Cronon:	could the position of the dart be stored at the time of collision?
[11:50]	Andrew Linden:	another bug fixed and waiting for 1.25-Server
[11:51]	Andrew Linden:	perhaps Rex, however you'd need a new LSL call to access that position.
[11:51]	Kooky Jetaime:	Thanks Orion
[11:51]	Andrew Linden:	What would you call it?
[11:51]	Orion Shamroy:	yw Kooky :)
[11:51]	Kitto Flora:	something like (vector) llDetectedContactPos( integer) ?
[11:52]	Rex Cronon:	u need to store the pos, and the key of object that it collided with, and the target needs to store the same info
[11:52]	Arawn Spitteler:	I'd suggest all collision data, since mass might change since. If the Object collided with isn' tphysical, it could change anything, and llSetPrimitiveParams has an inherrent delay, for other events to wait behind
[11:52]	Rex Cronon:	maybe for the last 16 collisions
[11:52]	Kooky Jetaime:	Orion - too bad they didn't do any really useful like 20x20x.250/20x20x.125
[11:53]	Orion Shamroy:	They bay Kookey... They had a ton of different sets over there, allmost all sizes.
[11:53]	Kooky Jetaime:	I made a 20x20x.5 so :)
[11:53]	Rex Cronon:	than from inside collision_end, u could reques that pos, like llDetectedColPos(int nr)
[11:53]	Arawn Spitteler:	Events wait for Server Lag, but also in Event Queue
[11:54]	Andrew Linden:	Well, it might be possible. I'll try to remember that when we start adding new features to the LSL API
[11:54]	Rex Cronon:	u already using an array that holds the latest 16 positions
[11:55]	Andrew Linden:	Rex, that is probably correct. We'd just have to store more info in those 16 slots.
[11:55]	Andrew Linden:	And then provide the new LSL call for getting at that info.
[11:55]	Arawn Spitteler:	You wouldn't know what a future scripter might find use for, and will have to store everything
[11:55]	Rex Cronon:	u could make a new struct that has an additional atttribute(pos of collision)
[11:56]	Andrew Linden:	No, we could store whatever we wanted to first expose, and then add more if necessary.
[11:56]	Andrew Linden:	Alright, this hour has another 5 minutes or so.
[11:57]	Andrew Linden:	Any last minute topics?
[11:57]	Rex Cronon:	we have free dessert here;)
[11:57]	Rex Cronon:	a cupcake:)
[11:57]	Kitto Flora nibbles Arawn's cherry
[11:57]	Andrew Linden:	With a cherry on top!
[11:57]	Arawn Spitteler:	llDetectedCollisionUV might be done with llDetectedTouchUV, but llEventUV would make simpler syntax
[11:57]	Arawn Spitteler:	The Deceitful One should plug SVC- 2931
[11:57]	Andrew Linden:	But UV implies texture coordinates
[11:58]	Orion Shamroy:	null calories = llAttackCupcake();
[11:58]	Andrew Linden:	many collisions are on corners and edges
[11:58]	Arawn Spitteler:	The Texture could carry collision targets.
[11:58]	Rex Cronon:	having the actual face, and the pos on it might be usefull, but more usefull is to know where the object was on collision
[11:58]	Arawn Spitteler:	EventFaceUV, with options for multiple faces.
[11:59]	Arawn Spitteler:	When the system detected the collision, since it could have teleported
[11:59]	Kitto Flora:	the hack-n-slash bunch would probably like to know where an Av got hit by a bullet
[11:59]	Andrew Linden:	Tell you what Rex, when it looks like we can start adding LSL changes again, I'll try to remember to remind you to open the jira item for that feature.
[12:00]	Arawn Spitteler's also curious about: Interest List, in Access controlled parcels, how infeasible is that?
[12:00]	Rex Cronon:	u just have 2 read the transcripts:)
[12:01]	Andrew Linden:	Arawn, please elaborate on that interest-list / parcel access idea
[12:01]	Arawn Spitteler:	When I try to visit a web-page, for which I haven't access, I'm not going to see anything. If I'm bouncing off an Access Line, I shouldn't see what's inside either
[12:01]	Andrew Linden:	You want it such that people who aren't allowed (or are explicitly banned) from a parcel to not be able to see object on that parcel?
[12:02]	Kooky Jetaime:	Andrew- that would be cool :)
[12:02]	Simon Linden:	I did some investigations / thought exercises on that ... called it "parcel privacy"
[12:02]	Andrew Linden:	And are you also talking about how people on that parcel shouldn't be able to see the banned avatars and their objects?
[12:03]	Rex Cronon:	can u ban everybody? including lindens?
[12:03]	Simon Linden:	I personally think it would be a good feature, but suspect there are a lot of edge cases that may be difficult to get right
[12:03]	Kooky Jetaime:	Andrew - I don't know, sounds interesting but I'd like to have that as an option, not compelled
[12:03]	Arawn Spitteler:	The privilages of the parcel privileged would be a related topic, but this is about privacy, rather than muting
[12:03]	Kitto Flora:	I've seen like 20 lindens on the ban list of a parcel :) I doubt it works...
[12:04]	Andrew Linden:	It does work... unless they go into admin mode
[12:04]	Rex Cronon:	maybe a linden is the owner and wants to ban other lindens:)
[12:04]	Kitto Flora:	Wasnt owne by a linden
[12:04]	Orion Shamroy:	Yay! Interoffice feuds. :)
[12:04]	Kitto Flora:	I never did find out why they were all on the list
[12:04]	Arawn Spitteler:	Are Linden Alts able to use Admin Mode?
[12:04]	Kooky Jetaime:	No Arawn
[12:05]	Kooky Jetaime:	Last I knew they can only do it when signed in as a Linden
[12:05]	Kooky Jetaime:	could you imagine the chaos?
[12:05]	Arawn Spitteler:	SuperNoob would be a fun game for G-Team
[12:05]	Andrew Linden:	ok, well I've got to go now. Maybe we can talk about the "parcel privacy" idea some more in the next hour
[12:05]	Kooky Jetaime:	It would make it very possible for "alts" to say I'm a linden, do it or I'll just admin and do it myself"
[12:05]	Andrew Linden:	We've had other Lindens ask for that feature
[12:05]	Arawn Spitteler'd also like a privilege list, so merchants don't have to fill up their groups
[12:05]	Simon Linden:	Me too ... thanks everyone for coming
[12:05]	Kooky Jetaime:	Not that people don't already try it
[12:06]	Kitto Flora:	Its a feature that would probably please many. True privacy within SL - at last?
[12:06]	Andrew Linden:	however there may be problems in trying to implement it...
[12:06]	Simon Linden:	Assuming we get all the chat, voice and updates sorted out :)
[12:06]	Kooky Jetaime:	Kitto - once Open Sim is more known, true privacy will be hosting your own region and holding discussions there
[12:06]	Andrew Linden:	I suspect some devs will want to overhaul the interest-list code before we started such a project.
[12:06]	Rex Cronon:	imaging somebody banning the grid monkies from entering:)
[12:06]	Kooky Jetaime:	s/known/mature
[12:06]	Kitto Flora:	Just what I was thinking, Kooky :)
[12:07]	Arawn Spitteler:	Ban all lindens, and invite them to a God-Mode Only party
[12:07]	Rex Cronon:	bye everybody
[12:07]	Kooky Jetaime:	There will have to still be a code of conduct for private sims, but
[12:07]	Simon Linden:	Bye
[12:07]	Kooky Jetaime:	you could eliminate most of the "rules"
[12:07]	Rex Cronon:	have fun
[12:07]	Kitto Flora:	Lunchy time!
[12:08]	Kitto Flora:	Byebye all
[12:08]	Andrew Linden:	Thanks for coming.
[12:08]	Orion Shamroy:	Live long and prosper Kitto :)