User:Ama Omega/archive/Office Hours/2010-08-23

From Second Life Wiki
Jump to navigation Jump to search

List of Attendees

Transcript

[09:07] Kelly Linden: Hello everyone. :)
[09:07] Melchizedek Blauvelt: and Kelly
[09:07] Penny Patton: "There's a giant box here with an SLURL...what does it mean?" Months later, a pile of skeletons is found sucking up all t he sim's spare prims.
[09:07] Kelly Linden: hehe
[09:07] Rex Cronon: hi Melchizedek, kelly
[09:08] Kelly Linden: So, the topic of the day, at least to start, is LSL calls for display names.
[09:08] Kelly Linden: NOT "is display names a good idea". :)
[09:08] Kelly Linden: The new LSL functions are on the wiki now, and I have a FAQ up: https://wiki.secondlife.com/wiki/Display_Names/LSL
[09:09] Click the seat to change your animation or position.
[09:10] Kelly Linden: Does anyone have any questions or concerns? Or is there another topic people would rather discuss?
[09:10] Jonathan Yap: Is there a date for when they will be up on aditi?
[09:10] Rex Cronon: how may words can the user name have?
[09:10] Kaluura Boa: Hello... I'm late... What did I miss?
[09:10] Kelly Linden: The username can not have spaces, the display name can be as many words as you can fit in 31 characters.
[09:10] Rex Cronon: what is the minimum length for user name?
[09:11] Jonathan Yap: Is a unicode letter 1 character or more?
[09:11] Rex Cronon: and maximum:)
[09:11] Kelly Linden: Jonathan Yap that is a good question I'm not sure about.
[09:11] Penny Patton: Oh wow. 31 characters long,
[09:12] Kelly Linden: The functions are ......
[09:12] Imaze Rhiano: so - only llGetDisplayName and llRequestDisplayName are going to return/use display names? everything else is going to use user names?
[09:12] Kelly Linden: live on agni. Except that the DisplayNames ones will return an empty string until the serviec is live.
[09:12] Kelly Linden: All existing functions will use legacy / full names.
[09:13] Kelly Linden: No existing functions will change to usernames. There will be llGetUsername and llRequestUsername for usernames
[09:13] Imaze Rhiano: display name doesn't need to be unique?
[09:13] Kelly Linden: Correct, it does not.
[09:13] Zep Palen: I am a little worried about the 72 hour delay on the displayname script functions
[09:14] Imaze Rhiano: any plans to add username to avatar uuid functions? :P
[09:14] Rex Cronon: what is the min/max length for user name?
[09:14] Kelly Linden: 72 is a bit ... extrenious. It should happen much quicker, but there is some cache propegation we are dealing with.
[09:14] Kelly Linden: rex: 1:31
[09:14] Zep Palen: But Kelly...with any delay it will give impersonaters a gap in where they can have a free play
[09:14] Imaze Rhiano: ya... going to fun write scripts for different contest - like photo contest when users change their name during contest :P
[09:15] Kelly Linden: For anything where you are tracking a specific user, use the key. If not the key, use the username or full name.
[09:15] Kelly Linden: Don't use the display name for tracking people.
[09:15] Zep Palen: with instante refelction it can be used to test for impersonators in shops...who would use the shop owners name
[09:15] Kelly Linden: Zep: Oh sorry, the 72 hour delay is not LSL specific
[09:16] Kelly Linden: That is how long it may take for your display name change to take effect.
[09:16] Kelly Linden: (this is in Torley's FAQ)
[09:16] Zep Palen: ok..so the change is the same for both LSL and the actual view in the viewers?
[09:16] Kelly Linden: yes.
[09:16] Zep Palen: ok...thats good...then i am happy ?
[09:17] Zep Palen: I have my questions answered...next victim
[09:18] Kelly Linden: Imaze: Actually the new names service that is the backend for display names does make it more feasible to create a username to key service.
[09:19] Kelly Linden: I think that will wait until display names have been out for a while first though. Let us iron out the feature and see the real world load and performance of the service first.
[09:20] Kelly Linden: Well that was far less exciting a topic than I had thought it might be.
[09:21] Sonya Haight: we can call a few more if you want more excitement
[09:21] Rex Cronon: if u would have aid that lltextbox is here:)
[09:21] Rex Cronon: have said*
[09:21] Zep Palen: I have a little addon to teh subject
[09:21] Kelly Linden: heh, use prim media!
[09:21] Kelly Linden: go ahead zep
[09:22] Zep Palen: sorry...thats more than 2 minutes ago..so I forgot it
[09:22] Kelly Linden: haha
[09:22] Kaluura Boa: Seriously, llTextBox() has been sitting on the depthes of the servers for how long, already?
[09:23] Kaluura Boa: in*
[09:23] Kelly Linden: I have a jira - SVC-5913 about http-request failing until a region restart - anyone here seen that?
[09:23] Jonathan Yap: lltextbox is on the list to be processes for the next software sprint
[09:23] Jonathan Yap: *processed
[09:23] Kaluura Boa: Wow! Revolution! hehehe
[09:23] Jonathan Yap: It was just discussed at Oz's OH
[09:23] Kelly Linden: prim media is so much more awesome though. :(
[09:23] Kaluura Boa: Except that it requires V2... So... No go!
[09:23] Kelly Linden: um
[09:24] Kelly Linden: What do you think lltextbox is going to require?
[09:24] Jonathan Yap: There is a lltextbox patch out -- it just needs converting
[09:24] Kaluura Boa: I'd like so much to play with it but I just can't stand V2 so I don't...
[09:24] Kelly Linden: any new view features that *we* do are going into viewer 2.
[09:24] Jonathan Yap: http://jira.secondlife.com/browse/SNOW-392
[09:24] Penny Patton: Sounds like changes are coming to the Viewer 2 UI, addressinga lot of the complaints.
[09:25] Kaluura Boa: As long as V2 is the way it is, anything V2-only is wasted time
[09:25] Zep Palen: llTextBox is a little more resource friendly I think for sits where you just need a user to add a line for something
[09:25] Imaze Rhiano: llTextBox is working in Emerald already - and I think in some other tpvs
[09:25] Jonathan Yap: It is also working in Cool Rainbow Viewer -- that patch applies directly to 1.22 base code
[09:26] Kelly Linden: right - if you just need it in TPVs then you have that. If you are hoping for somehow making it work with the 1.23's already out there that isn't going to happen.
[09:26] Kelly Linden: so getting it on our backlog just gets it into viewer 2 / snowstorm
[09:26] Rex Cronon: seems like teh code for llstextbox is already available. ll just has to get it:)
[09:26] Jonathan Yap: Rex, will you put it into Snowglobe 1.5?
[09:27] Rex Cronon: am i a linden?
[09:27] Kaluura Boa: Yes, you are... =^_^=
[09:27] Rex Cronon: nope
[09:27] Kelly Linden: the problem is going to be the same as prim media still - you will have to code multiple methods to use depending on what their viewer supports.
[09:28]
[09:29] Kaluura Boa: V2 is a serious bump under the scripters carpet... Where is my hammer?
[09:30] Kelly Linden: Are there any scripting issues specific to viewer 2?
[09:30] Imaze Rhiano: last time when I tried V2.1 - it's script's editor just crashed or stopped viewer several seconds when copy pasting larger scripts
[09:30] Kaluura Boa: The V2-only functions
[09:30] Jonathan Yap: Getting the script editor to know about new functions?
[09:30] Zep Palen: I remember my forgotten thing from before
[09:30] Detected name: Kelly Linden
[09:30] Username name: kelly.linden
[09:30]
[09:31] Kaluura Boa: Hoooo! Working already?
[09:31] Zep Palen: Would you concider changing the words on the FAQ about the display name lsl functiions..they are easy to misunderstand
[09:31] Ardy Lay: It's in Server 1.40
[09:31] Kelly Linden: Jonathan, ah. yeah that is a good one. We should download the keywords file from the server on login if it has changed.
[09:31] Kelly Linden: Zep: Which words?
[09:31] Zep Palen: quote; "Note: It may take up to 72 hours for a display name change to take effect, and scripts may report the old display name during this time."
[09:32] Kelly Linden: Display names will return an empty string though until the service is enabled.
[09:32] Zep Palen: it is easy to read as it goes for scripts only
[09:32] Kaluura Boa: 72 hours? The hamsters in the wheels are so tired?
[09:32] Rex Cronon: u guys change the names manually:)
[09:32] Kelly Linden: I think they are just giving themselves wiggle room. but I think we have a 24 hour cache on the names
[09:32] Kaluura Boa: Seriously? 72 hours in computer world, that's an eternity
[09:33] Zep Palen: as long as they change at the same time for both viewers and scripts then it is fine with me.
[09:33] Kelly Linden: Zep: It may take up to 72 hours for a display name change to take effect. During this time scripts may report the old display name and viewers may see the old display name.
[09:34] Kelly Linden: how is that?
[09:34] Zep Palen: yes
[09:34] Liisa Runo: we dont want scripts that "may" work
[09:34] Kelly Linden: I think there will be some fuzz around the end of the time when the change actually takes effect, but it should be really close.
[09:34] Zep Palen: perfect
[09:34] Kaluura Boa: For me, that makes sense... but it's really slow
[09:34] Rex Cronon: is kind of weird. 72 in a digital age to change a name?
[09:35] Kelly Linden: ok wiki updated
[09:35] Zep Palen: thanks
[09:35] Kelly Linden: It is really slow.
[09:35] Kelly Linden: I'm pretty sure it will be faster in reality.
[09:35] Zep Palen: and the delay in itself is actually good...it prevents som sudden impersonators with no patience to work
[09:36] Kelly Linden: I believe even sites like facebook can take a long time (in computer time) for a name change to propegate so others see it.
[09:36] Kaluura Boa: I think it would be better if you could change fast and then not again until a certain delay
[09:37] Zep Palen: there is already planned one change per week only
[09:37] Kelly Linden: You sort of get into a race and diminishing returns as you work to make the change faster and faster.
[09:37] Kelly Linden: But in the end we have viewers that cache names, simulators that cache names and central servers that serve the names also heavily caching the names
[09:38] Kaluura Boa: I'm not into it but I easily imagine some frustrated Role Players... "Let's go and play. Need to change my name. See you tomorrow or the day after... maybe"
[09:38] Rex Cronon: why so much caching, when max nr of users online is < 100k?
[09:39] Kelly Linden: We show names everywhere.
[09:40] Kaluura Boa: What about some dedicated server(s) that could be interrogated in real time?
[09:40] Kelly Linden: Reducing the time is something that could probably be done as a second phase or project if there is enough demand.
[09:40] Jonathan Yap: Just think of how there are a zillion dns calls being done every day
[09:41] Kelly Linden: Kalurra: There will be a dedicated 'names' server. It will sit behind a cache.
[09:42] Imaze Rhiano yawns... do we have other topics than these "display names"?
[09:42] Zep Palen: dns servers also got delays
[09:42] Kelly Linden: Imaze: I actually don't this week. Do you have one?
[09:42] Rex Cronon: imaze. make your own topic:)
[09:42] Kaluura Boa: What's in the pipelines? What new functions are coming?
[09:42] Imaze Rhiano: okay - how is mono rezzing fix coming?
[09:43] Kelly Linden: Oh! Good topic.
[09:43] Penny Patton: Ooo, Mono rez issue. Also curious about that.
[09:43] Kelly Linden: There will have to be 2 phases. We spent a bit of time trying to avoid that but it didn't pan out.
[09:43] Kelly Linden: Similar to how we had a mono-aware phase before mono was added, there will be a mono2-aware phase.
[09:44] Imaze Rhiano: okay - how are memory limits coming?
[09:44] Kelly Linden: We need this because current sims can not handle the new mono2 scripts.
[09:44] Kelly Linden: However, we believe we are very close to getting both phases ready for testing. Hopefully this week.
[09:44] Kelly Linden: Memory limits is not currently being worked on.
[09:45] Zep Palen: has it been dropped?
[09:45] Kelly Linden: Not dropped. I suspect it will start getting more resources "soon" - though I'm not sure of an exact time line.
[09:45] Morgaine Dinova: Kelly: do you mean that memory limits are not in the scrum backlog, or they're not even in the product backlog?
[09:46] Morgaine Dinova: (Oz told us today about the distinction)
[09:46] Kelly Linden: They are not on the current sprint for the performance team.
[09:46] Kelly Linden: But they are on the performance teams backlog.
[09:46] Morgaine Dinova: kk
[09:46] Penny Patton: From Philip's keynote, sounded like he was keen to introduce performance boosting limits.
[09:47] Kelly Linden: He is, but there are pre-reqs to script limits that we have to do, and they still get weighed against other backlog items even if philip is keen on them.
[09:47] Kelly Linden: Believe it or not the mono sim freeze work is one pre-req.
[09:47] Penny Patton: Heh. Was just saying it's unlikely they were dropped. Bit of chat lag.
[09:47] Kelly Linden: "small scripts" is another soft pre-req.
[09:47] Kelly Linden: Ah. yea.
[09:48] Zep Palen: you mean dynamic mem assignment?
[09:48] Kelly Linden: allowing Mono scripts to set a lower memory ceiling.
[09:49] Zep Palen: yes..thats what I meant
[09:49] Kelly Linden: Then yes. :)
[09:49] Zep Palen: that in itself will also release resources
[09:50] Imaze Rhiano: any plans to open development process like snowstorm scrum team did?
[09:51] Kelly Linden: Jumping topics: I talked with andrew Friday about link distance rules. He is going to try a new approach to calculating the bounding sphere (which is one very expensive part of the calculations)
[09:51] Kelly Linden: Maybe he can make a significant enough change.
[09:51] Kelly Linden: Imaze: no.
[09:51] Kaluura Boa: Yeah... We talked about strictly distance based limits
[09:51] Rex Cronon: now we will be able to link anything inside a 64m radius:)
[09:52] Kelly Linden: We are hesitant to do anything that would allow really large link distances until we have better parcel/region border handling and collisions across region borders.
[09:52] Kaluura Boa: That should be much faster than the complex formula nobody understands
[09:52] Kelly Linden: Rex: ah cool.
[09:53] Rex Cronon: i am hopping that will be the min link distance:)
[09:53] Kelly Linden: Well, we could get away with 55m I think.
[09:54] Kaluura Boa: 57m... Weird number
[09:54] Zep Palen: hmmmm isnt 55 the max limit ow?
[09:54] Zep Palen: now
[09:54] Kaluura Boa: 57
[09:54] Kelly Linden: the current limit is 54m I think, and that is a bounding sphere including the size of the objects being linked
[09:54] Kaluura Boa: Or 54...
[09:54] Rex Cronon: now we have weird formula that computes the link distance:(
[09:55] Kaluura Boa: I have doubts
[09:55] Kelly Linden: The code says 54.
[09:55] Kaluura Boa: So be it! Make it possible to link anything within a radius of 54m and we'd be happy... for a while
[09:56] Rex Cronon: at t least 64m:)
[09:56] Liisa Runo: 64 is nice round number
[09:56] Zep Palen: So is 100...
[09:56] Kelly Linden: hehe. I still kind of like the 'w/in 55m of the root prim' idea, but 110m diameter objects makes andrew want to improve our large-object handling
[09:57] Imaze Rhiano: ooo... does that means he is finally ready for megaprims?
[09:57] Kaluura Boa: LEt's hope!
[09:57] Kelly Linden: No, it means once we get to that size he wants to fix the megaprim issues before we do it.
[09:58] Imaze Rhiano: "fix the megaprim issue" ? :P
[09:58] Kaluura Boa: Encroachment
[09:58] Kelly Linden: s
[09:58] Kelly Linden: issues
[09:58] Kelly Linden: encroachment and apparently also allowing things to colide across region borders
[09:59] Rex Cronon: and to think that detecting ecroachment is so easy..
[09:59] Sonya Haight: no offsim mountains?
[09:59] Kaluura Boa: Some ppl will be very upset...
[09:59] Jonathan Yap: I found a bug with off sim rocks
[09:59] Kelly Linden: you can always set the object to phantom
[09:59] Zep Palen: does that mean I have to remove my 2048x2048 meter offs im landscape? ?
[09:59] Liisa Runo: and when the megaprim issues have fixed. LL can easily give us megaprims up to debug_channel. all kind or talk about silly 30m prims makes me think of soaked duct tape
[09:59] Sonya Haight: my waves! don't take my waves!
[10:00] Kelly Linden: That is a good thing to bring up to andrew, zep.
[10:00] Zep Palen: i dont dare that..he'll kick my butt
[10:00] Kelly Linden: haha
[10:01] Kelly Linden: It would be awesome to have better skybox support
[10:01] Penny Patton: Definitely worth bringing up. Anything that can be used to make SL look so much nicer shouldn't get tossed without consideration.
[10:01] Kaluura Boa: I'd like the skybody not to be a box...
[10:01] Penny Patton: Ooo, better skyboxes would be nice. Would love to see the night sky get a long overdue improvement.
[10:01] Kaluura Boa: skybox*
[10:01] Liisa Runo: lights that go trough everything would be awesome
[10:01] Liisa Runo: dont go trought *
[10:01] Kelly Linden: I think PRIM_ROT_LOCAL is code complete for some upcomming (post 1.42) server release too
[10:02] Kelly Linden: And with that, I have to run! Thanks all.

Generated with SLog Wikifier