Firestorm meeting 13th February, 2013 – video and transcript

Update February 25th: As per a comment from Jessica Lyon, Firestorm have now merged the Server-side Baking code and updates to RLVa into one of their private repositories.

firestorm-logoOn the 13th February, 2013, the Firestorm team hosted a question-and-answer session so they could outline the current status of the Firestorm viewer, the issues the team (and SL) are facing, and outline plans for both the immediate future and longer-term as well as address questions from the audience.

While the meeting was recorded, the Firestorm team are aware that many of their users have hearing difficulties, and / or prefer to read text. It is because of this that this transcript has been provided, otgether with the original recording itself. When reading it, please remember:

  • This is not a word-for-word transcript of the entire meeting. While all quotes given are as they are spoken in the video, to assist in readability and maintain the flow of conversation, not all asides, jokes, interruptions, etc., have been included in the text presented here
  • If there are any sizeable gaps in comments from a speaker which resulted from asides, questions to other speakers or requests for images to be displayed, these are indicated by the use of “…”.
  • Timestamps are provided as guidance should anyone wish to hear the comments in full from any speaker on the video
  • Questions were asked in chat during the meeting and while speakers were talking. This inevitably meant that replies to questions would lag well behind when they were orginially asked. Therefore, to provide context for both questions and answers, questions have been included in the transcript and timestamped at the point at which each is addressed by a member of the Firestorm team
  • The meeting proper commences at 16:36 into the video, and this is the point at which the transcript starts.

Video courtesy of Northspring.


16:36 Jessica Lyon (JL): OK, I think we’re pretty good to get started folks. So the way we’re going to do this is kind-of like an informal thing, so I’ll probably talk for about 15 – 20 minutes about where we are, where we’re going, what the future is looking like, what challenges we’re facing, blah, blah, blah … and I’ll ask you guys if you can hold your questions until we’re done on that, and once we’ve done that, we’ll field your questions in local chat. We’re not going to do a whole “send you questions to so-and-so”, because it just gets really complicated and goes over time; so that’s what we’re going to do.

17:32 JL: So a lot of you have heard about Server-side Baking, which is coming. Now we haven’t quite gotten quite to the point yet of merging-in Server-side Baking into our repro … Tankmaster [Finesmith] is dealing with out merges; we’ve just recently merged-up to Linden Lab TIP, which is their most recent release – what version is that Tank?

18:14 Tankmaster Finesmith (TF): 3.4.5

18:15 JL: OK, so we’re merged-up to Linden Lab 3.4.5 locally, you know, internally. And … I’m going to be completely honest with you guys. Right now Firestorm, for us internally, is in pretty bad shape since our merge with Linden Lab’s TIP. There’s a lot of bugs that we’ve inherited; there’s a lot of regressions which we’ve inherited. Ed [Merryman, lead for Firestorm Support] is crashing about two times a day – and for those of you know Ed, know that Ed never crashes. So if Ed is crashing on our recent builds, we’ve got some problems. We’ve got some log-out crashes, log-out things; log-in crashes … Basically, we’re not in great shape, and we’ve got a lot of fixing-up to do before we’re ready for a release. But I’ve had a lot of people saying to me, “What we really need is a release.” Well, we can give you what we have right now, and I can tell you right now that it is way worse than our last official release. So you do not want a release from us right now.  So this is before we even merge Server-side Baking into our code …

19:50 JL: So I’m starting off with the bad news. There is some good news, but I’m starting out with the bad news. So Tank has his own private repository, where he’s been merging-in Server-side Baking … and this needs to be released in a timely manner; in fact Linden Lab is quite concerned that third-party viewers can get Server-side baking released in time for them.

20:22 JL: So this puts us in a bit of a spot because right now, even before we merge Server-side Baking, our current repository is in some pretty bad shape and it needs a lot of TLC, and that’s going to take us some time to fix. Then we’re going to merge Server-side Baking in, and that’s going to give us a whole bunch of new bugs which we’ll also need to fix.

20:46 JL: Some of the things which are holding back Server-side Baking right now are RLVa, which is significantly broken by Server-side Baking. Kitty’s been working on that recently, poor Kitty – Kitty Barnett. As most of you know, Kitty does most of the RLVa, and Server-side Baking has a pretty heavy impact on RLVa functionality, so that needs a lot of work. And then there’s going to be a lot of testing need to make sure that that’s working. Unfortunately, Linden Lab has only given us a couple of regions on the beta Grid [Aditi] – which, I may add, is not very stable right now – and there are things that are stopping Kitty from being able to test RLVa on those Server-side Baking regions. For example, scripts are disabled there, so she can’t really even do the testing. I sent in a request yesterday [February 12th] to turn on scripts on the Server-side Baking regions, and I’m sure they’’ do that because they really need Server-side Baking out[scripting was enabled on one of the regions on the 15th February].

22:01 JL: But the bottom line is that we’re being pushed into a corner in a way. We have to get Server-side Baking merged-in with our code and there’s going to come a point where we’re going to have to release what we have – and what we release is quite possibly not going to be pretty. And that’s not just us; all third-party viewers are having or are going to be having problems, I think, especially the ones who have merged-up with Linden Lab with their latest release. And we’ve been pulling-in some of the bug that we inherited; we’ve been pulling-in from their beta in the hopes that that might help fix some of the problems – it’s a pretty long list, so I’m not going to go into all of the details of all of the bugs that we have.

22:56 JL: The worst of it is that when Server-side Baking does get rolled out, if you do not update to a viewer that has Server-side Baking that’s what you will see

The SSB problem in part: I'm stabding on an SSB-enabled region. On the left  - as I appear to others who are using an SSB-enabled viewer; On the right, as I appear to others who are using a viewer which does not support SSB.

The SSB problem in part: I’m standing on an SSB-enabled region. On the left – as I appear to others who are using an SSB-enabled viewer; On the right, as I appear to others who are using a viewer which does not support SSB. (Note – this is not the image shown in the meeting; the image shown in the meeting relates to a SSB failure resulting in an incorrect COF update, an entirely separate – and apparently resolved – issue)

23:21 JL: So the worst news is really is our next release, for the most part, is probably going to suck as far as stability is concerned and we’re probably going to have a lot of bugs, and y’all aren’t going to have a choice, you’re going to have to use it or you’re going to be stuck with that.

23:42 JL: I can give you my word that we’re going to do everything in our power to try to stabilise Firestorm and get bugs and kinks out before we release it, as much as we humanly can. But like I say, there is a timeline on this, and we don’t quite know when that is going to be. Linden Lab gave us a two-month warning, [but] it looks like that’s going to be a little bit longer now because Linden Lab has the same bugs we just adopted from them; but the bottom line is it’s going to go out, and we’ll have to release Firestorm before we’re satisfied with it – and that really hurts me a bit, because I hold our devs up to a high standard with Firestorm. Firestorm has had the …  lowest crash rate … for something like four or five months straight, something like that, maybe even more – a long time, anyways, and I’m really proud of how stable Firestorm is. And now a lot of you are going to be saying, “But I crash all the time.”

24:54 JL: The crash rate is an average of all of our users, and how many of those sessions ended in crashes. Some of you may be crashing constantly and your crashes do have an impact on our crash rate, but it’s an average of a lot of users, and we have a lot of users, and currently our crash is … I think it is eight percent, just over eight percent, which is extremely low for Second Life standards.  So I’m pretty proud of that, but I expect that’s probably going to change with the next release.

25:33 JL: Now, on top of that, on top of Server-side Baking, Linden Lab is also merging-in CHUI [the Communications Hub User Interface]. CHUI is some interface changes  – for the most part, interface changes on the Linden viewer – and which, for the most part, we’re not too concerned about, because we kind-of like our interface. There’s a couple of things in CHUI which we might like to cherry-pick. And out initial plan was to go through, “Hey I like what they did with chiclets! Let’s grab that and pull that in!” Unfortunately, Linden Lab mixed the interface changes with a whole lot of code refactoring and changes … which we do need to bring in to Firestorm.

26:56 JL: So what that’s going to mean is that we’re going to have to merge-in … well over a thousand commits. We’re going to have to merge-in all of this stuff to our repository, and because it also contains a lot of interface changes, it is going to horribly break Firestorm.

27:20 JL: So then what we’re going to need to do, is go in and try to fix everything that is broken. And really, when you get to a stage like that, the amount of testing which has to be done is ridiculous, because Firestorm has so many options and features and interface changes and skins  – and everything has to be meticulously tested, so that’s going to be devastating for us.

27:48 JL: On top of that, we’ve also got the materials project that’s coming in to the code, and we’re going to have to merge that code, although that probably won’t be too bad, I don’t think.

28:22 JL: We’ve recently been merging-up with the Linden Lab HTTP changes that have been coming, and there’s still quite a bit more to do server-side, from what I understand, on HTTP.  So the moral of the story and the point I’m trying to get to is that the next couple of months, I think, is going to be very difficult for everybody. Not just our team, but I think for all third-party viewers. Even Linden Lab. I have a feeling, are going to have a tough time server-side with things as well, and basically, Second Life, I think we can safely say will be largely FUBAR for probably a few months.

29:09 JL: The good news from all that is that it’s all for a good cause, because all this stuff which is coming is really cool. Like Server-die Baking, once it’s working well and stabilised … you can expect that your avatar will rez … avatars will rez almost immediately when you teleport into a region. HTTP, if things work out the way it is expected, you’ll have less teleport failures, things should rez faster … I mean in a perfect world, if everything works out right, all these changes are hopefully going to be worth the months of headache we’re all going to have moving forward here.

29:56 JL: So 2013 is kind-of looking to me like it’s going to have a kind-of shaky start, I think, and hopefully a really good end in the mid-to-last part of the year, so there’s that.

30:12 JL: Good news also is that we do have some pretty cool features coming with Firestorm … You know how we always recommend you guys always do a clean install, that you wipe your settings and do a clean install, and I know how much you guys hate to do that? We stated, I think it was in the last Office Hours, that we really want to not have to recommend you guys do that in the next release.

30:40 JL: So what Zi Ree has done now, is that we have an import / export working for your settings. So  you can actually save all your settings  into a folder on your computer somewhere  – and if you want, you can even put that on a USB stick and take that if you go to your friend’s house or somewhere else or to your laptop, you can transfer those settings, and you can import them back into the next installation. Really, really cool stuff, and we’re hoping beyond hope that will end the whole problem of having corrupt settings and we’re also hoping that will eliminate the need for you guys to be doing clean installs or losing your settings all the time. I realise how much of a pain that is for people, as we have to do it as well, especially when we’re testing things and so we do it a lot more often than you guys do, actually, and so we know exactly how much of a pain that is. When you have a viewer with as many features as Firestorm has, and options, trying to go in a fix everything to how you had it is next to impossible, and takes weeks sometimes, because you forget things. So there’s that.

The user settings back-up Preferences tab - coming to Firestorm 4.4.0

The user settings back-up Preferences tab – coming to Firestorm 4.4.0

31:53 JL: We’ve got – and I don’t know if that’s going to be ready in time [for the next release] – but import / export is coming. Techwolf [Lupindo] has been busy – I mean import / export of prims, your content, basically, if you make content. That’s being written from scratch, it’s being really well written and we’re hoping that other third-party viewers that do import / export will use what we’re putting out as a standard for importing / exporting. This is being designed to work in OpenSim as well as in Second Life, and it’s designed to work very well and its extremely strict with permissions to prevent any kind of abuse as much as possible, and so that’s coming.

33:07 JL: We put out Quick Preferences, and most of you probably know what that is; it’s a little button that you can enable and inside of it it has, you know, your draw distance and your particles – basically the things that we thought most people would want in their Quick Prefs. But as it turns out, everybody is different, some people want different options in there than others, and so, you know, we have people saying, “Well, why isn’t this is there?” and, “Why isn’t that is there?” And, you know, we try to find the best balance for everybody.

33:37 JL: But Quick Preferences now lets you put whatever options you want, pretty much, in there. So it’s customisable. Quite literally, you can change the options that are in it … There are some limits; do you want to talk about it, Ed?

34:30 Ed Merryman (EM): Just basically, anything that’s a debug setting, you’ll be able to add to your Quick Prefs is the basic limitation.

34:43 JL: So we expect that to be hugely popular, because …  people who have combat, they have different things they want to change, and this lets you pretty much customise whatever you want, within certain limits … and I believe they do get stored, they do get saved in the settings export.

35:30 JL: We’ve got a whole bunch of really cool stuff that’s in the works. Unfortunately not as much as we usually have on most of our releases, but that’s because we’re pretty much swamped with bugs that we’ve adopted with the merge  – and I’m not trying to say that all the bugs we have are Linden Lab’s fault; in many cases we’ll have bugs which happened because of the merge or as a result of the merge, and some of the bugs we have because the Linden Lab viewer has them, and some of them are our own. So we’re working pretty hard trying to fix those for you, so a lot of people, like I say – and I’m just trying to wrap up here so we can go to questions – a lot of people are saying, “Give us a release! Give us a release!” – Right now, you do not want a release from us, take my word for it. We need to get Server-side Baking in as soon as possible. That is our top priority, because if we don’t and Linden Lab release baking on the server-side … you guys, you know, ”Avatars, everything’s broken, it’s terrible!”

36:40 JL: So it has to be our first priority; normally stability is our first priority, [but] in this case that has to be our first priority and stability has to come second. So our next release may be suffering from stability. I hope not, I can be wrong; I was expecting a higher crash rate in our last release than we actually resulted in having. You know, we were pleasantly surprised by that. But either way, we’re crashing, and if Ed is crashing twice a day then things are probably not very good. So we’re working on it, but we’ve got problems with tcmalloc again, we’ve got, like I say, just all kinds of problems. Can you guys think of anything I’m forgetting?

37:27 EM: Well, there’s the option to hide chiclets; there will be some preferences which get renamed because of certain changes which are being made by Linden Lab, and there’s a bunch of little things that we don’t think about very often.

37:46 JL: Well actually the chiclets thing isn’t really a small thing. All you’ve got to do is read the blog comments on our blog and you know people hate chiclets – and most of us hate them to, although most of us have, I think, gotten used to them now and we just don’t even notice them any more. So … Pantera[polnocy Resident] has done some coding that allows you to hide most of the chiclets. Not all of them, unfortunately.

38:16 JL: You know, there are limitations to what we can do, folks, we’re not magicians, and the v3 code is not as flexible as we would like it to be. So there’s a lot of things that are just difficult for us. For example …  another one of our goals was to … option out the ability to eliminate chiclets and bring to you the option to have v1 notices, group notices, script dialogues, behaviour back.

38:51 JL: We’re only going to be able to give you half of that, so we’re giving you the option to remove chiclets, and that’s going to remove all chiclets, I believe, except for inventory offers, group notices … and it’s not for lack of trying, really. We would love to bring back the v1 dialogues and script notification dialogues and behaviour – and it is still on our list of things that we still really want to do, because it drives us nuts to; like the whole thing about script dialogues cascading down off the screen to the right, you know, well that’s ridiculous. We feel that; we know that, we hear what you’re saying and we totally agree with you. It’s just not as easy to do as it is to say, “Can you do this?” So it is on the list.

39:49 JL: Also, v1-search, or legacy search … Cinder has incorporated the legacy search in the same floater as the web search. That has been a much tougher job than we expected it to be, so kudos to Cinder for that, and a lot of changes to that. If and when LL turn off legacy search, it’s gone; you’ll only be able to use web search, in which case … the next build we release just wouldn’t have it. But OpenSim does use legacy search, so we would keep it, certainly, for OpenSim.

Legacy search functionality is being extended and integrates with the v3 web-based search

Legacy search functionality is being extended and integrates with the v3 web-based search

41:06 EM: And as Ansa[riel Hiller] pointed out,  standalone windows for your teleport history, landmarks, place details … standalone window for the block and mute list. For the people that use the classic vintage skin, you know how you’re got those music controls both on your bottom bar and up on the menu bar? Well you’ll be able to get rid of the ones on your menu bar if you like. That’s an example of the little things I’m talking about, Jess.

41:51 JL: So basically it’s good news and it’s bad news … the bad news is that we’ve got some stability problems and we’re on a time clock that is sort-of beyond our control, so … at some point we will have to release what we have, whether we like it or not … you know, I have this thing where I have a hard time standing behind a product that I don’t believe in, and I feel I’m going to have a difficult time with this next release because I really don’t like giving you guys a bad release.

42:38 Question from Innula Zenovka: Is it really the case LL intend to release SSB as soon as it’s ready, even if that means breaking Firestorm and RLVa, rather than wait a few weeks until Firestorm and RLVa are ready? I am hot questioning what you say, but I do find it a bit surprising they’d break every TPV (and that’s an awful lot of products.)

42:50 JL: I can say pretty safely that Linden Lab is really, really hoping that we’re going to get Server-side Baking out and I honestly don’t believe they’ll push it out until we have it ready. BUT … there’s a limit to that. We can’t just drag it on and drag it on and drag it on because we’re not stable, just give us another month, that’s not going to happen; at some point there’s going to be a deadline and we’re going to have to push out what we have.  And that also depends  – it’s not just Firestorm, there are other viewers; Singularity had quite a few users … and so Linden Lab can’t really roll Server-side Baking out until it’s largely available in most of the third-party viewers, because as you see by that picture, it really does kind-of break everything.  But there is going to come a point where management is going to say, “Enough is enough.  Roll it out on the server.”

43:50 JL: So I don’t think they are depending on us to get it out, but they are really hoping to, and I’ll tell you they’ve been very, very helpful in offering help; Nyx Linden has been fantastic, Oz [Linden] has been really good … every chance they get, they say, “If we can help, just let us know.” Unfortunately, they can’t help us with things like RLVa and they can’t really help us with some of our stability bugs.

44:28 Question from Hugsalot Valkyrie: Are you guys going to push back the release of Firestorm after LL implements server baking?

44:35 JL: No, we’ll have to have it out. Firestorm will have to have Server-side Baking released, stable or not … we’re going to have to have something out; we can’t leave you guys no option; you’re going to have to have at least some Firestorm viewer you can get on to see the world properly.

45:26: TF: No viewer out now will work correctly after SSB is released.

45:52 Question from Hugsalot Valkyrie: is LL pushing back the Server-side Baking? Or is LL going to keep client-side baking while pushing out server-side?

JL: There is no in-between. It’s an on/off switch, so there’s no backwards compatibility. Once they turn it on in the server, it’s on. And you need to be on a viewer that can run it.

47:25 Question from Crim Mip: Is there any way that the OpenSim version of Firestorm could install to and use a different location to the non-OS version? I’d like to be able to have both versions installed.

48:03 Cinder Roxley (CR) in chat: It’s definitely planned for the next release. Top priority for me. [One of the biggest problems has been separating] cache

48:09 Question from Ilana Debevec: Have you considered an interim release to handle some of the more obnoxious bugs (FIRE-8456)? Ed may be crashing 2 times a day, I’ve been crashing 8-10 times a day and I like my HEAD!

EM: One thing you have to understand; I’m crashing twice a day on the nightlies that we’re testing internally. If I come in on the release, I’ll crash maybe once or twice a month. That’s just an indication of how the nightlies are, presently. Will we do an interim release? You’re probably going to end up with an interim release that includes the Server-side Baking, and that’s what it’s going to be, because we’re probably still going to have … a lot of bugs, most likely in this release.

49:08 JL: In other words, we could give you an interim release [right now], and we can pretty much guarantee you it’s going to be 10 times worse than what you’re on right now, unfortunately.

49:53 Question from Goldvald Enoch: Will Firestorm be releasing a viewer that is able to bake both on OpenSim and Second Life, or would you need one viewer for Second Life and on for Open Sims?

50:04 TF: The short answer is yes. So with the Server-side Baking code [in the viewer] it will bake with old and the new service; it’s not like you have to be on the Server-side Baking server in order to bake. It will still use the old way if it detects that you’re on a region that still has the old code.

50:27 JL: In other words, we’re not deleting the – let’s call it the legacy baking code; we’re not deleting that from Firestorm. So basically, when you log-in to a grid, whether Second Life or an OpenSim, it’s going to check to see which grid you’re connecting to [actually, check the region] and say, “OK, I’m going to use this code to do baking,” or if it’s, say, OpenSim, “I’ll use this old code then, to do the baking.” And so the viewer will be backwards-compatible, but the Linden server won’t be.  So when Linden Lab flips the switch on the server, the old baking code just won’t get used [on the server-side] … so in that way you won’t have to have a separate build.

51:20 TF: Just as a little bit of a note,  on the old clients, you’ll still look fine to yourself because you’re still baking locally. However, the bake that your viewer makes will be sent to the server, but the server will ignore it … if you’re editing your appearance on the Server-side Baking code, it will still render you locally for the duration of the editing, then it will send the changes to the server why you hit Apply, and then the [new] server bakes, so the service isn’t going to be baking every time you’re making small changes as you’re editing yourself.

52:14 JL: At least, that’s as best we understand it right now – and we reserve the right to be wrong! Linden Lab changes their mind frequently, so do we, so what we state right now could change next week or tomorrow. But this is as we understand it right now.

52:37 Question from Bob Paderborn: With group notices, if you say unchecking disable Notices will this be fixed?

JL: That is a pet peeve of mine, and I don’t think it has been fixed. This is one of those, “We can’t figure it out” bugs? … So, Bob, maybe, maybe not … I feel your pain, though. I’m owner in lost of our support groups and most of them are languages I don’t speak, and every time I log in, they pop-up for me, so I’ve gone in and set them all up so they don’t pop-up chat for me, and it seems like two or three logs I’ll be fine, and then the fourth log suddenly they’re all popping up again, and it seems to have forgotten that I’ve set that setting.  So, yeah. I feel your pain.

53:48 Question from Jewelyette Resident: Have you already talked about the stream change probs at entertainment sims

CR in chat: I fixed that and it will be in the next release.

54:21 Question from Suicideblonde Scheflo: Any hope of having option to stop group notices from spamming that group’s already opened chat window?

54:44 TF: Right now the behaviour is still the same.

54:56 EM: Your only option is to is to really disable group notices for that group.

55:03 Question from Adham DeCuir: What will be some positive things about server-side bake? Things that you are looking forward to?

JL: I am looking forward to never having to worry about whether I appear correctly to everybody else. Bake fail is a very real thing and I think every single person in Second Life has experienced bake fail at least once, and I would venture to say more likely one out of every ten times they’ve logged in they’ve got a bake fail. So … the whole reason for this re-write of code is to address the bake fail problem. So if things go well, and it works out the way Linden Lab expects and hopes it to work out, you won’t have to worry about whether you’re naked to somebody else while you look dressed to yourself. You shouldn’t have to worry about having two or three different outfits being worn when to you, you look fine, or you look broken to yourself but you look fine to everybody else. Bake fail has all these different variations, and that’s probably the biggest thing I’m looking forward to. Also the speed of having avatars render quick would be nice as well.

56:47 Question from Volmarrgoth Volmar: After this craziness has passed will there at some point be a 64-bit version of Firestorm?

JL: People have been asking for a 64-bit Linux because it is possible, people have been asking for a 64-bit Mac because it’s theoretically possible. I can say that both of those are in our long-term plans. We talked briefly about having an experimental option for people who want to live on the edge with 64-bit. So that’s a possibility, but we’ve got to get through this mess first, I think.

57:28 Techwolf Lupindo (TL): It is on my long “to-do” list, to get 64-bit builds out for Linux. I’ve been trying to do that for the last couple of years.

57:44 JL: And Tonya [Souther] wants to work on a Mac 64-bit as well. And … the next obvious question is “What about windows?” Well, that might be a different story; we have to look into that, but it’s far enough down the road it’s not even worth talking too much about yet other than to say we’d like to at some point.

58:08 Question from Nalates Urriah: Phoenix 1185 had shape export to OBJ file. Will that EVER come back to FS? It is the only free way to get shapes into Blender.

JL: That’s been requested a lot, and we’ve talked about it a few times. And it is absolutely on our list of things to do. Probably I would expect to have that within the next couple of releases of Firestorm. So it will come back.

58:33 Ansariel Hiller (AH): It actually works, but can’t be released due to license issues. Mech OBJ export has from Singularity to work with mesh and needs some relicensing and things.

58:35 TF: I think Nyx said he was going to look into it, something that would bring back that capability into the Linden Lab viewer.

59:08 JL: Oh, Goldvald points out that Server-side Baking is going to kill temporary textures, folks. So if you use temp textures, that will be broken when they flip the switch. You do, though, also have the local textures

1:00:15 Question from Jewelyette Resident: I find that LMs attached to toasts go stale instantly. Is there a way to fix that?

1:00:23 EM: There is a workaround for that … In your group notices, if you open a group notice through that toast that you get, if you open the group info through there, you’ll find that you can’t get the landmarks a lot of the time, or the attachments. However, if you open it through your Contacts list, you’ll find that you can still get them.

1:01:00 TF: What’s happening is the server drops the attachment after an hour of it not being clicked on.

1:01:10: Kadah Coba (KC): Is this related to the ancient bug due to the design of where the inventory offer from the notification expires?  Because I know that’s been a [LL] bug from the beginning

1:01:23 EM: I think it is

1:02:07 Question from Trin Trevellion: since some months, many ppl see avatars in wrong sizes. Usually the shape is too short or too tall especially at crowded areas. Is there any more information about this bug.

JL: There absolutely is. Whirly [Fizzle] was talking about this the other day, it is if I remember correctly a server-side issue, and there’s a Linden JIRA on it.

1:03:40 Question from Hugsalot Valkyrie: Any new fonts or skins?

1:03:51 JL:  We have another skin coming out

1:03:51 CR in chat: Ectoplasma theme for anastorm.

1:04:00 CR in chat: User-defined cloud normals. I should know, I wrote it, and Vincent Nacon contributed his cloud normal maps and a mass of his windlight settings specially made for them.

1:04:22 EM: There are some additional windlights that have been added that are – awesome, I guess is the only word for it.

1:04:46 TF: Cloud normals isn’t just a new windlight theme or whatever, it allows you to overlay another texture over the sky to give it more definition … so you get a little bit better 3D effect.

1:05:45 Comment from Hugsalot Valkyrie: I heard something about new bump mapping functions in SL a couple weeks ago….

JL: That’s the materials project … the Materials Project is pretty cool … it brings a new dimension to texture in Second Life with the caveat that you have to have deferred rendering enabled. Actually, there’s something I need to talk about with you guys…

With and without: how materials will look when running a viewer in differed mode (top) and in non-deiffered mode (bottom). The differences are clear, but the in-world experience in non-differred mode is not in any way "broken"

With and without: how materials will look when running a viewer in differed mode (top) and in non-deiffered mode (bottom). The differences are clear, but the in-world experience in non-differred mode is not in any way “broken”

1:06:45 JL: You know how when you go to your Graphics Preferences you’ve got Low, Medium, High Ultra? Linden Lab has expanded those options; so now there’s Low, there’s low-medium, there’s Medium, medium-high, High, high-ultra and Ultra. And they’ve done this in preparation for the materials project that coming … So first off this caused us some grief internally, because we don’t necessarily agree with Linden Lab’s defaults and we don’t agree with each other in what we think should be the defaults. So we’ve had an ongoing sort-off a battle internally with deciding what are the best defaults, where is the best place to have deferred rendering enabled as a default … So basically we concluded that we weren’t going to turn deferred on as a default until you reach high-ultra …. You know when you log-in to a new viewer and it says, “We’ve detected your graphics card and set the graphics accordingly”, that’s what I’m referring to when I say, “default”. So Linden deferred comes on in their viewer at medium-high.

1:08:35 JL: So a lot of people are going to log-in to the Linden viewer, and suddenly they’ll have deferred. And deferred … one of the pet peeves of mine … many of you may still be wearing shoed that use invisiprims, like myself, and when you enable deferred, your feet show through your shoes; so that’s annoying.  And what’s going to happen is, a lot of people are going to have deferred on by default now when they first log-in, and certainly I wasn’t OK with that as a default for Firestorm users.  So we did a big vote, and we’ve decided that we’re turning deferred on as a default at the high-ultra setting, and no matter how good your card is, it will not be defaulted to anything higher than high.

1:09:35 JL: So if you want to see the materials project, the materials work, you’re going to have to click deferred rendering in the Graphics options to enable that. And it’s going to be called Advanced Lighting Model; that was Linden Lab’s terminology. Now  I’ll say the materials project is pretty cool actually, but it does required deferred rendering and that does come with some consequences for invisprims and it also comes with a performance hit for some people, so there’s that.

Updated Graphics tab options reflecting LL-deriven GPU table updates and upcoming support of materials processing

Updated Graphics tab options reflecting LL-deriven GPU table updates and upcoming support of materials processing

1:10:19 TL: Just to give everybody and idea, I recently experienced something similar to what Jess was explaining about. When I rebuilt my Firestorm to the latest TIP which had these changes in there and I didn’t know about, I logged in and I used to typically get between 30 and 60 frames per second, depending on whether I’m on a crowded region or whatever, and I logged-in to the same place I’d just logged out from, and I was getting ten frames per second, and I knew right away that something was wrong. And further investigation revealed that what happened was that my graphics card was detected and it reset the default settings and I got that message at start-up and I didn’t think much of it, and what it did was, instead of putting me at setting which gave me similar performance level … it bumped me up two notches … and gave me that horrible performance. And as soon as I knocked that back to two notches down, I got my performance back. But my experience, I know, will reflect what will happen to a lot of people , because like Jessica said, a lot of people will log-in like me and say, “What the hell happened here?!” And that’s going to be a problem , and we’re just going to have to get the word out that Linden Lab made some changes which are not going to be obvious to fix.

1:12:07 Lette Ponier (LP): And if you don’t have Tech’s experience, you may have my experience. Where I’ve always logged-in with a default on High on a clean install and now a get on to Medium, so your new default settings may be some place completely different to what you’re used to, so you’re either going to be bumped-up and you have worse performance than you’re accustomed to, or you may be bumped down and maybe your feelings will be hurt.

1:12:45 JL: My biggest concern was: perception is everything, right? It’s something we’ve learned through doing this whole viewer thing … and if normally you don’t run with deferred rendering on the last build and you update to the new build and suddenly deferred is enabled for you – most people don’t know what deferred is – suddenly you discover you’re getting like 15 frames per second less than the previous build, the reaction of the user is going to be, “This version sucks! I’m going back to the old version!” And really, it’s just because that you didn’t know that deferred has been enabled by default for you. So that was one of my arguments why we shouldn’t enable it by default. But there’s a lot of arguments both ways.

1:13:25 TF: Just a note: deferred is Lighting and Shadows in the Graphics Preferences.

1:14:32 EM: Jess, before we start losing people, I’d like to plug the classes a little bit … For those of you who don’t know, we do have classes pretty much every day on various aspects of the viewer and some things to do with Second Life. A lot of people actually know that, but only the people who actually come out to the classes know that we do a Q&A session after every class, where you can ask anything you want … I’d like to encourage you to let your friends know, to come out to the classes. People who come out to these classes learn things and as a general rule, I think that anyone who has come to the classes will point out that they have fewer problems  because they now have a very clear idea of what the viewer is and what the Preferences actually do.

1:15:44 JL: I find it interesting actually, because … I often see in blog comments people saying. “Why should we have to go to classes just to learn the viewer? Firestorm sucks, blah, blah, blah.” You don’t have to come to classes.  Our whole mission statement is to improve the user experience, and we do that also with classes. If you want to improve your experience, come to classes. You don’t have to come to the class, but I can guarantee you, you’re going to learn something that you didn’t know, and you’re going to say, “Wow, that’s cool now I know that that does that!” And I know this because I learn stuff when I go to the classes still. And you know, Firestorm has a lot of stuff, a lot of options and things, and no you shouldn’t have to go to a class, and no you don’t have to go to a class, but if you do go to a class, I can guarantee your experience is going to be improved because you’re going to learn stuff.

1:16:38 LP: Even the instructors learn at most classes. It’s rare that I don’t teach a class and come away either learning something from the other students or noticing something new in the viewer that just never struck me, I just never noticed before … And Willow makes a great point: teacher learn how people use the viewer in different ways. That’s actually the way we learn the most, by direct interaction. S yeah, lots of reasons to come to classes.

1:18:45 JL: So how are Phoenix and Firestorm comparing with each other? Unfortunately, I can’t give you exact numbers, but I can say that there’s over twice as many people on Firestorm than there are on Phoenix. I can also say that nine out of ten Phoenix people switched to Firestorm rather than another viewer. So we are losing some, and honestly, and I’ve always said this. If Singularity or Cool VL or some other viewer improves your experience better than our viewer, then I want you to go to that viewer. Really I do. But we’ve been pretty fortunate, most people are switching to Firestorm , they give it a few days, and they love it. It’s just giving it a chance … yes, you’ve got to learn some stuff, things are moved, the menus at the top are different, but it really doesn’t take that long to learn – a couple of days … I do have stats, I wish I could share them. I can say Firestorm, by a long way, is the most popular viewer right now.

1:20:11 TF: It’s about equal for V3 and Phoenix.

1:20:15 JL: Yeah, Phoenix and the Linden V3 viewer are pretty neck-and-neck right now. Phoenix is about to drop below the Linden viewer in usage … and it goes Firestorm, by quite a bit, and then Phoenix and then just below Phoenix is the Linden viewer, and every week Phoenix is going down, so it will drop below the Linden viewer before too long.

1:22:17 Comment by Cigarette Girl: I love to hate it. I love the UI, but it cooks my Mac

1:22:19 JL: That’s a  problem that we’ve been trying to identify and it exists as well in the Linden viewer … that happens to some people and it’s very difficult to reproduce as it either happens to you or it doesn’t, and it happens to a lot less people than it doesn’t happen to.

1:23:05 Question from Hugsalot Valkyrie: So did someone take over old phoenix viewer?

JL: Funny thing about that. I’ve had people say, “Why are you being selfish and killing Phoenix?” and I’ve have people say, “Why don’t you let someone else take it over?” And we’ve stated many times that it is free to take over – and in fact there have been a couple of people who have attempted to take it over, and failed. And I’ll tell you why they failed. Phoenix needs a crap load of work to get it back up-to-date and it needs a heck of a lot more work to try to keep it up-to-date. And the way I saw this is: anybody who is smart enough to successfully recover Phoenix and renew it, and get it up-to-date – anybody smart enough to do that is smart enough to know it’s more work than it’s worth; it really is. And so I’m not surprised that the people who have attempted it have failed and have actually moved on to other viewers instead.

1:24:04 LP: Personal plug. There’s been some commentary actually on the prospect of Phoenix being adopted by somebody else. I wrote in my own blog about that, I think Tonya did to, actually …

1:24:54 Question from Dove Redyard: I’m finding differences between Phoenix and Firestorm in abilities to copy and paste textures while building and wondering if it is meant that way or if it is a bug?

JL: There are differences, obviously, between Phoenix and Firestorm and there are more changes coming, even in regard to the build floater which are not ours but are rather Linden Lab’s. And we’ve had complaints; some people hate it, some people like it, and I don’t know if we’ve made any plans on making changes to that. It’s going to boil down to it’s a learning curve; it’s just something you’ll hopefully get used to.

1:25:48 Question from Hugsalot Valkyrie: Is there a date from LL when server side baking is pushed out?

JL: No. And your guess is as good as ours – and in fact our guess is as good as Linden Lab’s. I don’t think they even know, because they’ve got a lot of bugs, just like we’ve got in Firestorm, they’ve got in their viewer and they got them on the server-side and they’re trying to work through them. And we’re sort-of hoping it takes them longer to work through them, because that gives us more time to try to fix our bugs. So all I know is it’s not on schedule, so it’s already behind.

1:26:50 Comment from Widget Whitebury:  I use Phoenix when I need to listen to live music. With FS, sound always stutters. I’ve tried recommended fixes from FS support

1:27:17 LP:  That’s often network bandwidth / connection kind-of stuff, especially if iTune’s doing it to.

1:28:36 Comment from Vollmarrgoth Volmar: Wishes he could get streaming to work on Linux. I have 64-bit Linux and Firestorm is 32-bit so it gets confused due to Gstreamer being 64-bit on my system and don’t think I can install 32-bit Gstreamer as well.

JL: OK, I’m going to be the villain here, just sort-of getting back to the 64-bit builds. I’m the bad guy because I am the person who is not in favour of 64-bit Linux or Mac, and I’ll tell you why.  Because as soon as we put out a 64-bit Linux, even though it may be buggy, there are certain things that don’t work, I understand, and even though we may get a 64-bit Mac out, which is even further down the road, but let’s make the assumption that we do, I think Linux and Mac together are 12 or 15 percent of all the users, something like that. So then we’ve got 85 percent of our users saying, “Where’s out 64-bit Windows build?” And again, perception is everything, and if there’s a 64-bit Linux and a 64-bit Mac, people are going to want 64-bit Windows even though it won’t actually benefit you anything. We’ve got SSE2, we’re taking advantage of your extra memory, and that’s pretty much what a 64-bit thing will give you. But I lost that fight. I do herd cats and occasionally the cats herd me, and I lost that fight, and we’re probably going to have a 64-bit Linux and a 64-bit Mac at some point.

1:30:09 TL: There have been attempts to release a 64-bit Linux, a few other third-party viewer have. If you’re a Gentoo user, you can use my overlay and build it on your system with 64-bit; I’ve had that for years, but you have to use Gentoo. There’s a lot of people that run Linux that run 64-bit, but we have just not released a 64-bit version because we then have to officially support it and all its annoyances. The biggest problem with Linux right now is you have about three major distributions, and you have to support all three of those … Unofficial releases are out there, and I’ve been half-tempted throw some releases out there for some of our users …

1:31:11 JL: We’re going to have a place on the wiki somewhere that’s going to be labelled “Experimental builds” and you’ll be able to get them there. Because although I lost that fight, I still will not put it an as official release, not unless it’s work really, really well.

1:32:13 Question from 1stlifescorpio Aftermath: When I first TP into certain sims (and this is prob 50% of the time) my mic automatically turns on w/o any visible notification that my local voice has been enabled and my voice is audible. This may have been happening to me in Phoenix.

JL: Yes, we are aware of that bug; we have no idea what’s causing it. We do know that it’s happening on the Linden viewer as well, are pretty sure it’s a Linden bug, and possibly even a server bug. We’ve not been able to reproduce it, and so it’s pretty hard to figure out what’s causing it. But yeah, it’s happened even during our office hours where somebody’s mic gets turned on and they don’t know it, and they’re talking away. It’s pretty serious, really.

1:33:16 Question from 1stlifescorpio Aftermath: What is the issue with support for “Restore to Last Position” breaking a few months ago and not being fixed?

LP: That was, I believe, a Linden server-side change that they have no interest in fixing because it only affects third-party viewers.

1:33:33 JL: If I remember correctly … that was a fix because griefers were able to rez objects in a region which has no rez, by rezzing them at 0,0,0. And so linden Lab made a fix for that which ultimately broke “Restore to Last Position”.

1:34:00 TF: It’s actually more broken in the Server-side Baking.

1:34:10 TL: If I remember correctly, wasn’t “Restore to Last Position” originally a Linden Lab feature?

1:34:17 TF: Yes, and then they never exposed it, but the server does support it.

1:34:27 TL: If memory serves me correctly, because I remember reading all the JIRAs about it, it was due to content loss. It turned out that people that exposed it and were using it lost their content half the time.

1:34:40 JL: It would go off-sim.

Top of Page

1:34:43 TL: It would go off-sim and be gone forever, be unrecoverable. It was a content loss bug that was being filed in JIRAs. And because of that, I can understand why Linden Lab just killed that feature.

1:34:55: Comment from Vollmarrgoth Volmar: My Firestorm runs at the edge of the 4-gig limit often and then crashes due to the memory limit of Windows 32-bit

JL: Actually, to be honest, if you were running 6 gig and it could use all 6 gig, it’s going to use all 6 gig and you’re going to crash because you ran out of memory. It’s another bug that we’re trying to deal with, and have been trying to deal with for quite a while now. It’s another one we picked-up, we believe, from Linden Lab, because they have the same problem.

1:35:26 JL: If I’m making Linden Lab sound like we get all our bugs from them, I apologise. That’s not true; it just maybe sounding that way. Believe me, we have plenty of our own bugs. But we combine our bugs with Linden bugs when we merge Linden code.

1:35:43 Question from 1stlifescorpio Aftermath: Is Purify or similar used to minimise memory leaks?

1:35:55 JL: Memory leaks are a huge challenge in the viewer. And the reason in part as to why it’s a huge challenge is because there’s more than one thing that’s leaking. So while we may find one memory leak, there’s a thousand more that we’ve got to find somewhere. And sometimes we have luck and sometimes we don’t.

1:36:34 Question from 1stlifescorpio Aftermath: Will “Restore to Last Position” ever be re-supported?

JL: No. Certainly not with Linden Lab’s viewer, and it just doesn’t work because they’ve pretty much broken it server-side while trying to fix something else, while fixing an exploit … I mean if you had to weigh it on a scale, what’s more important: Keep “Restore to Last Position” or keep letting griefers attack help islands and things by putting out prim rezzers, I would certainly be in favour of Linden’s decision on that. I’d sooner eliminate that exploit than save that feature. But it is missed.

1:37:22 EM: Although you can still use it if you have rez rights at 0,0,0.

1:37:27 JL: Yeah, that’s right. If you have rez rights at that particular position in the region, “Restore to Last Position” will work. So it’s not that it’s completely broken … you have to have permission on that pert of the region in order for it to work. But there’s probably no hope for it coming back.  [JIRA: https://jira.secondlife.com/browse/SVC-7907]

1:39:46 Question from Hugsalot Valkyrie: are system minimums specs going up for FS?

JL: If they are going up for the Linden Viewer, they are going to go up for Firestorm. That’s a really tough thing, because we have people who are running one gig of RAM on Windows XP, not even a service pack, and they’re running it fine. And then we have people running Windows 7 with 8 gig or RAM, and they can’t run it. So it’s really hard to come up with a minimum system requirements. We can do a recommended, I suppose – and we’re going to recommend you have a really good computer!  But pretty much whatever Linden specs are, that’s what ours are.

1:40:53 Comment from Trin Trevellion: you should convince LL to make something about that look-at cross-hairs or at least they should make it optional this is a drama feature.

EM: I happen to agree with you in one aspect. However, we teach in our class that that lookat doesn’t actually show where your camera is. It show’s the camera focus point. I could focus on anybody in here and then be perve-camming anybody else I want simply by moving my camera around. You also have the option in Firestorm to turn that off, so you’re not showing your camera focus point to anyone – and it does make a very small difference on server-side lag, because it’s less information that the viewer’s pulling from the region that you’re on.

1:41:49 JL: Let me add to that.  Was it about a year ago? … they discovered that these lookat beacons use up a lot of server bandwidth, and they were at the time trying to figure out how to lessen that. And so a Linden was looking at eliminating these lookat beacons, and they had quickly come to us asking questions about it, because we’ve got this option in the viewer for it. And he didn’t quite understand that the lookat beacons are actually necessary in order for your avatar to look animated. With default settings on Firestorm, your eyes look around, they follow your mouse a little bit. And if you disable these lookat beacons,  you become focused like a little zombie kind-of thing. So basically, Linden Lab was going to remove the crosshairs altogether until they discovered that that actually degrades the user experience, and so they decided the gain wasn’t worth the loss. And that’s the reality of it, that feature is there and it can be enabled through any viewer through debug settings, and so its not that we made the feature; it’s inherent in how your avatar is animated.

1:43:50 JL: And it is one of those features, though, that has some negative use cases and it has some positive use cases. Like using it like a laser pointer. Often a lot of people who do teaching will use it as a laser pointer to point at a certain prim that they’re talking about. When I was a mentor … one of the tricks we did so that students knew what we were talking about, and which prim we wanted them to look at, we would tell them how to turn on your show lookat beacons so that they could see my lookat beacon sitting on the prim that I want them to be looking at while I go and torture the prim or do whatever it is I’m doing. So there are some good use cases for that, and in fact it was those use cases which is why it was enabled in the first place. Unfortunately it gets used in bad ways as well.

1:48:00 EM: Angel’s question is about the different skins, and if I understand it correctly, rather than having them all included with the package, she’d like to see them downloadable like they were on Phoenix … and we have actually discussed that internally and as Tank says, it’s on the “to do” list.

1:48:36 Question from Adham DeCuir: What are some “feature requests” that you have looked over that you really want to implement but cannot yet?

JL: Obviously, as we mentioned earlier, the whole v1 dialogues and notices behaviour … I don’t think anybody on this team are happy about the way Firestorm and V3 in general handles script dialogues and group notices and whatnot. We would love to change that, but it’s turning out to be really, really difficult to do. So that’s a biggie, certainly for me and I think for a lot of us and certainly for a lot of users who have asked for it.

1:49:20 JL: A lot of people have asked for default prim parameters. In Phoenix you’d be able to set when you create a new prim it has this texture by default and looks this size, is this shape, or whatever. A lot of people have asked for that. But we can’t give you that until we have import / export done. The good news is that’s being worked on.

1:49:48 EM: We have a default size, the X, Y, and Z parameters and we have a default material, we just don’t have the texture yet.

1:50:05 TL: There were some in there, I think there were two or three parameters which I did use in my re-write, but the rest of it I have to add.

1:50:25 JL: So the long story short of that is that we’re hoping that once Tech gets the bugs worked out with import / export, that we’ll also be able to provide that option.

1:50:38 JL: Other things … killing chiclets with a flamethrower. But again, that’s proved to be a lot more difficult than it should be. And another common one is, “How come we can’t reproduce the v1 interface?” We’ve got the Phoenix mode in Firestorm, but no it’s not exactly the same as the Phoenix viewer or v1 or 1.23. And we would love to be able to provide you that exact look. But again, that’s just one of those things where it’s a lot harder to do than it is to say. A lot harder. In some cases there are things that are virtually impossible to reproduce behaviour or appearance. There’s a lot; there’s quite a bit of it actually.

1:51:37 TL: You can also interpret what Jess is saying as a challenge to any developer that’s out there. If you can manage to pull that off, you’d be highly welcome.

1:51:49 EM: I for one would like to see the animation explorer in Firestorm.

1:51:57 JL: Although we’re not stalled from doing that, it’s just a matter of somebody doing it.

1:52:06 EM: There’s a tonne of things. Everybody’s got their own their own little thing they’d like to see in the viewer, quite honestly. Or things.

1:52:26: Question from Hugsalot Valkyrie: So CHUI replaces FUI?

CHUI: The new conversations floater in it's three modes (open, collapsed, minimised) - Linden LAb viewer

CHUI: Includes many updates and refactors and a new approach to communications. Shown here is the new conversations floater in it’s three modes (open, collapsed, minimised) – Linden LAb viewer

JL: So CHUI is Linden Lab saying, “We heard your complaints, and we listened and we made changes to our interface that you didn’t like.” Kind of like FUI was; people complained about the sidebar, so they got rid of the sidebar, and everyone’s yelling and screaming because the sidebar’s gone. CHUI is kind-of like the same thing; there are things people hated. Chiclets – people hated chiclets – so Linden Lab did something with the chiclets. We’re not sure we like what they did with the chiclets. They obviously have the same problem that we do in getting rid of the chiclets, and what they’ve done is put the chiclets in its own little floater in CHUI; its got its own little window which you can move around. Which is cool, if people like to move it around, we’ll try to give you the ability to move it around. But even moving the chiclet bar to another part of the viewer screen is really complicated and hard to do in this viewer code. So a lot of people are saying, “I don’t want chiclets on the bottom right I want them on the bottom left, or on the top left.” … Believe it or not, that is a lot easier to say than it is to do; even though we were able to move it to the bottom right, it’s a lot harder to move it to the left. And you kind-off have to be a coder to … understand why it’s so hard.

1:54:03 TL: And I can personally attest to that.  When I was doing my area search re-write, I dug into Linden Lab’s UI code to add a certain feature to it that was requested  and it proved to be quite difficult because of the way the code was. I ended up adding a feature to linden Lab’s UI code that half-way works. And unfortunately, that’s all I code do: make it half-way work, and that was it. The code was just that ingrained, that spaghetti-like that I couldn’t make it feature complete.

1:54:47 EM: And I’m going to say this, but I’m going to preface it with two comments. One, this is my own personal opinion, not the team opinion; and two: I’m really not a Linden Lab hater,  but CHUI is Linden Lab’s way of saying, “Well, we’ve heard there are things about FUI you don’t like. So we’re going to change it, and we’re going to give you other things that you don’t like, so you’ve got something else to complain about.”

1:54:18 JL: Some of it is OK. But … I think our UI is better, and I think our usage statistics indicate it’s better. CHUI – in fact FUI – resembled the Firestorm interface changes, and CHUI also does. And that’s cool; I like that we’ve inspired Linden Lab with some things.

1:55:52 Question from 1stlifescorpio Aftermath: When i exit FS, is FS flushing the entire cache to local disk? Sometimes it takes a long time to exit gracefully from FS. is there anything i can do to accelerate the exit time?

JL: I can’t give you technicals about that, but I think some of our devs can, and we call it a log-out hang … and in some cases the viewer starts to flash white and black. I’m not sure we even know what causes that yet.

1:56:38 TF: The cache is saved as soon as the pile comes in; it’s not like saving it to memory and then saving it to disk.

1:56:59 JL: It’s an exit hang, yeah. Just like when you log-in, there’s all kinds of communications going on between you and the Linden server. When you log out, there’s all kinds of stuff going on there as well. Communications that are being severed and whatnot and … sometimes the viewer is waiting for the server to tell it something, and it doesn’t come, and the viewer will just sit there and wait and eventually you just disconnect.

1:57:33 Comment from Vollmarrgoth Volmar: Why doesn’t LL adopt Firestorm improvements into their viewer?

JL: Well the sort-of have. A lot of the changes they’ve done resemble Firestorm in some ways. I believe we have given them some contributions; but I think Linden Lab, a lot of the time, likes to be the author of their own changes.

1:58:28 Question from Hugsalot Valkyrie: I’ve noticed on my Mac when I crash it asks me to do a crash report to Apple, not FS?

JL: That should come to us – and we do read them. Whirly will attest that she reads them, and Nicky [Dasmijn – Firestorm Lead developer].

1:59:48 Arrehn Oberlander (AO): We avoid some kinds of changes because they make us less compatible and harder to stay in sync. So we avoid some changes LL can make very easily.

1:59:50 JL: That’s a good point and I should probably iterate on that. There are some things we can do to, for example, make Firestorm look more like Phoenix. A good example of this, and it caused some tension among the team in fact, is the upper menu bar. I had the thought that there’s a lot of people coming over the Firestorm from Phoenix and they hate the fact that it’s all foreign up there … it’s not the same from Phoenix to Firestorm. So I asked Wolf Loonie, “Can you do a switch? Make it optional, so that people can choose to have the menu system of v1 rather than the menu system of V3. “ … But there are a few things. First, from a support perspective, that’s a nightmare, because … when you go to support and say “I’ve got this problem…” and they know that it’s just a setting that you’ve changed and they can just say, “Go to the build menu, select Link Parts…” If you’ve got a different menu system running, they can’t do that and that makes their life really hard. So that was one of the biggest arguments against it.

2:01:19 JL: And the other thing is, somebody has to maintain that. Every change that we make from Linden Lab, every time we move away from the Linden Lab base, every time we make a new option or a new choice, especially with interface stuff, eventually we have to merge-in a bit of Linden code, and suddenly everything that we’ve changed becomes a conflict during that merge. So the more we change, the more difficult it is for us to maintain Firestorm with the Linden code;  the harder it is for us to keep up with Linden Lab. So while there is a lot more stuff we could do, we have to balance what we can do with what we should do. We have to be responsible and we have to try to think forward. We have to think ahead.

2:02:32: Question from Hugsalot Valkyrie): do you literally make a new build every night?

EM: Yeah, one of the devs actually does a build every night for each of the operating systems.

2:02:40 JL: And you don’t want it. Trust me.

2:03:00 JL: What our nightlies are, is basically a compile of the code that we did that day. And there are day where Cinder, for example – and I’m just using Cinder as an example – will be working on feature X. And she’s got half of it done, and she’ll commit it to the repository –it’s only half finished, and she’ll commit it to the repository. And she’s going to do the other half tomorrow, and commit the other half tomorrow. Well, the nightly picked up half of that feature, and so that is broken. And if we started giving people nightlies …? Yeah … because Cinder is just one dev, and we’ve got a bunch of devs all working on stuff and all committing things every day. And so nightlies are more-or-less a way for us to track our progress.

2:04:26 JL: So yeah, there’s a lot of people ask us why don’t we have nightlies, and in fact I’ve got a few devs on the team who feel  we should release nightlies to the public … Only over my dead body!  That’s one fight I’m not going to lose … unless I’m no longer here!

2:04:53 TL: A perfect example of how we unintentionally break someone else’s build is that when we first did the internal test of the export / import, it turns out that it was broken on Windows. And my investigation revealed that Windows does things differently in ever-so-simple ways, but just enough to break a feature.

2:05:22 JL: There’s at least three stages to our quality assurance. The first is we’ve got our nightlies. And nightlies are almost guaranteed to be broken every night. Every single time it’s almost guaranteed to have some really nasty bugs in it. And it’s completely unpredictable because we don’t know what it’s going to do until we get it in our hands and run it. And then we discover that this is broken, that is broken and the other thing is broken.  That is not going to go into our beta testing. In fact that doesn’t even pass our internal testing. So there’s alpha testing.

2:05:55 JL: Alpha testing is where we give it to our support people – “Here, can you run this and let us know how it goes?” … If support comes back, and our devs come back and say, “This nightly isn’t all that bad, actually.” And we’re at a point where we’re due to put a beta out to our beta testers, to our QA, we say basically on a loose basis, “I think it passes internal.”

2:06:30 JL: And then the beta testers get it, and they find all kinds of stuff that’s wrong with it. And then they come back to us, and they say, “This is horribly broken,”  and so then we’ve got to fix this, that and the other thing. And then that build goes back to our internal testing with our support people until it passed that; then it might go back to the beta testers, and then they tell us all these things are wrong … and this is why it takes us a long time to get releases out. Because we’re pretty strict on bugs; we want to make sure that everything works as well as possible.  Because generally, if we have a bug which affects two people out of 200 beta testers, that’s 1%. Well, one percent of the amount of people we have using Firestorm is a lot of people. So if we release with that bug, there’s going to be a lot of people with that problem, and our support are going to be banging on my door, screaming at me for having a release with that bug. So we’re pretty quality conscientious.

2:07:53: One of the key parts of our QA … is Whirly. I do not know how Whirly can do it, but she’ll find bugs before we even have a build to test. She’ll somehow use here psychic abilities to know this  is going to work, and sure enough, it’s broken. So Whirly kind-of comes before the alpha. Whirly is like pre-alpha. I don’t know how Whirly does that stuff.

2:08:21 EM: Frankly, I don’t care how she does that stuff. I just love that she does that stuff for us

2:08:32 JL: Whirly is a human supercomputer … But basically, we just want to be sure that we give you the best release we possibly can. And it’s very disappointing, honestly, it’s actually very disappointing and discouraging when we go through all this process over and over with our quality testing and all these things, and finally we’re really sure we’ve got a pretty good release, we push it out, and we watch the support groups. And when we see people, “Oh, this isn’t working and that isn’t working..” … Every release, for me, anyway… is a happy and then kind-of a sad moment. Because it goes out and we’re really happy, really proud of all the work we’ve done, and it’s bittersweet because people are going to have problems. We can’t make it work for everybody, and that can be depressing and discouraging.

2:09:40 JL: Some of you may still be in our Firestorm Preview group.  [For] those of you who have never heard of it, when we first started working on Firestorm, we started up a group called the Firestorm Preview group. And this was for basically getting the feel for what people thought about what we were doing; it was just brand-new and we were just touching it, starting to work on it. That group is still around, although we’ve disabled chat and for the most part, it’s a dead group, although there are still quite a few in it. I think, once we get through, maybe even before we get through, this mess … I’m thinking about re-opening that group and using that group as a release candidate group. In other words, once we’ve passed the beta testers, we’ll throw it out to the people in that group as a release candidate and we can get a feel, because there’s going to be more people in that group. And we may be able to get a feel if there’s something we missed.

2:10:46 JL: I’m not going to give you a date on that; I’m not going to say, “Yes, for sure we’re going to do it”, but probably we will. I think it’s a good idea, but it still concerns me a little bit because we’re throwing out a potentially broken viewer.  And the other problem of course is that we’re going to say, “Please keep this to yourselves,” then 90% of the people are going to give the link to there friends, and their friends are going to give the link to their friends … and before you know it, we’ve got a release out, and then people in the support group are saying, “Oh. There’s a been a new release?” But there’s not been a new release, “But how come they are on a newer version than I am? And it’s another can of worms; so everything comes with good and bad.

2:11:30 EM: And then support’s going to tell you, “Sorry, we don’t support that nightly,” and people are going to get upset – but there’s just no way we van do support for a nightly, folks. There’s just absolutely no way.

2:11:50 JL: I think we’re actually going to close-up here, we’ve been a good two hours now … I want to thank all of you for coming; you guys are our favourite kind of users, and what makes you our favourite kind of users is because you care enough to come here and learn and ask questions and interact with us. And it’s because of this interaction we get from you guys, that we know what you guys want, and what bothers you and what doesn’t, and what we’re doing right and what we’re doing wrong. And we should do more Q&As,  and I think we will do more Q&As, maybe some more Office Hours and try to get back into interacting with you guys a little bit more than we have been – that’s largely my fault, real life has been really crazy. But thank you all for coming … and thank you Pantera for recording this …

2:13:11 JL: Yeah, we really need you guys as help to help educate people about Server-side Baking. Unfortunately, even our blog only reaches about 2% of our users. So there’s going to be a lot of people who have no idea what’s going on when Server-side Baking rolls out. So, you’re all in different groups, you see people screaming in a group, “This isn’t working, that isn’t working!” – you understand why, and you’ll be able to help them.

2:13:49 JL: Thank you everybody, and thank you, the devs and support folks who came out as well.

Related Links

About these ads

17 thoughts on “Firestorm meeting 13th February, 2013 – video and transcript

      1. Inara Pey Post author

        Have added the link into the two Related Links sections and the body of the transcript. Thanks, Ansariel :).

        Reply
  1. Pingback: Q&A Results « Firestorm Viewer – The Phoenix Firestorm Project Inc.

  2. garvie

    Isn’t anyone besides me a little concerned that an annoyance or a pet peeve dating from 2011 when sl adopted openGL3 is seemingly the logic behind moving a setting currently required for seeing materials in SL in 2013 is getting bundled off in to the “ultra” graphics settings? I refer to comments from 1:06:45 – 1:09:35. Also saying if you want to see the “materials project”, makes it sound as though materials are a temporary feature or something, which is not my understanding at all. I see them as a very welcome much anticipated and permanent new feature of content creation in SL, though obviously optional rather than required.

    I guess I can see a TPV deciding to use different default start up settings than the LL viewer for any number of reasons or even on the advice of an astrologer, but taking what the LL (sort of a default itself) viewer uses as a medium high default, (which on my set up is the same as what they used to call “medium” and moving that off to “high ultra” strikes me as a bit more than merely choosing a different set of defaults. In fact it strikes me as flat out misleading and irresponsible unless the FS team knows something about the terrible costs of running deferred rendering, why not just leave it in medium or medium high?

    I was hoping that we could at least potentially start using normal maps without having to put in a note card explaining why the product you just rezzed might not look nearly as nice as the picture you saw in the ad. If this note card must also explain that the setting the end user needs to apply will be in different places on different viewers, and that some viewers might have even have re-branded as ULTRA HIGH what other viewers will treat as commonplace, then I don’t see this boding well for materials being welcomed as the prim savers and realism enhancers they can be.
    Sorry for that run on sentence.

    In closing please, people QUESTION the wisdom of this sort of move and what it means for the community as a community that might one day see the same stuff the same way and even maybe the way the person who created it hoped they would. Meantime, you can detach and delete your invisi-prims, open GL3 rendered that ‘project’ obsolete a while ago.

    Reply
    1. Inara Pey Post author

      I can see both sides of the discussion – so to take the flip side to your comments…

      Many people in SL are, as Firestorm state, entirely unaware of deferred rendering, what it does and what it means (e.g. in terms of non-functional invisiprims). I’m not talking relatively new users either – I need both hands and my feet to count the number of people of 4+ years experience in SL as to the problem of invisiprims and deferred rendering – as they all still wear boots /shoes which rely on invisiprins and couldn’t understand why people kept nudging them about “broken shoes”.

      Materials aside, running the viewer in deferred mode comes at a performance cost. How ig a performance cost really depends on a wide variety of things – and not always purely the GPU someone has (I’ve had comments passed in this blog that my Ge9800GT running a 5+ year-old PC gets far better frame rates in deferred than more recent PCs with more capable GPUs. As a result, and in addition to the above:

      • There are those, who even aware of deferred, much prefer to run with it off, wither because their system can’t handle it very well or they simply prefer things that way, as they are involved in combat, etc., where deferred can be a performance detractor even on a reasonably good computer.
      • There is is a reasonable number of people running SL on computers whicah can’t actually handle deferred without grinding to a halt.

      When all of these factors are taken into account, and even with the best will in the world in trying to communicate to everyone what materials actually means and what is required to see the benefits of materials once it starts to make it presence felt – particularly given that so few people actually read blogs, be tween LL’s owen, the Firestorm blog, this one, or any other reporting on SL news – then it is very possible a lot of people are going to get upset for one of two reasons:

      • Their viewer suddenly isn’t running as well as the previous version and (becaseu they have no awareness of deferred and have never looked at their graphics settings) they are angry and upset and demanding that the “bug” is fixed or the release rolled back
      • They get annoyed because, even though they understand what deferred is and what it does, they actually don’t want it on by default, and get annoyed because they now have to disabled after every install.

      While these might sound like trivial matters, they can rapdily add-up in terms of support calls, etc.

      Ergo, one can actually understand the Firestorm team erring on the side of caution initially – particularly given that materials, with the best will in the world, isn’t going to transform the look of SL overnight – or even in a couple of months. It will take time. Plus, there is no reason why, as people become better acquainted with materials as the usage spreads, the Firestorm team cannot adjust their default settings to be more in-line with LLs.

      For my part, as both a Firestorm user and someone keen to play more with materials, I didn’t really have an issue with Jessica moving on the side of caution.

      Reply
      1. Ansariel Hiller

        Deferred not only causes a perfomance hit, but in it’s current state it opens a can of new bugs that are all long-known to LL but haven’t been fixed since.

        Personally, the most annoying bug is that LL didn’t just only break the invisiprims, they also break partially transparent textures in the form of colors being calculated wrong if they get hit by projector lights. If you don’t know what I’m talking about, have a look at this picture: http://www.flickr.com/photos/ansariel/8492749359 While the transition of the foot to the lower leg looks perfectly okay in non-deferred mode, you will look just as somebody wearing invisiprims: plain stupid! And this is going to happen to a lot of content, especially content using the currently highly used blending textures for seamless transition of mesh body parts – for instance the shoes, mesh hands, other “bits”…

        This is a common issue in deferred rendering and there are actually ways to mitigate this effect. Also this bug has been reported to LL almost 1.5 years ago until it was finally closed as expected behavior due to performance reasons: https://jira.secondlife.com/browse/SH-2638

        While the materials project itself is really nice, the bugs in the current state of the deferred rendering pipeline renders it pretty much useless. Unfortunately nobody seem to care at least until now. Maybe it changes if more people will suddenly find themself looking stupid – or somebody will rise some awareness of the issue… /me looks at Inara

        Reply
        1. Inara Pey Post author

          Yeah, aware of the other issues, and have poked about some of them. confined myself to invisiprims as Jess mentioned them in the meeting, so it’s a point of contact / context with her comments :).

          Not sure – from a lay user’s perspective (as opposed from apurely technical standpoint) – bugs in deferred rendering leave it “useless”; “problematic”, yes, but many do actually get by with it in the state it is in. But the fact there are issues is a valid point.

          As to raising awareness … with whom? If LL, then I can always shuffle along an d make room for you on the bench at the Opendev meetings :).

          Reply
  3. garvie

    i get the ‘caution’ approach, but to me erring on the side of caution would mean stopping short of making an ultra graphics settings requirement for what other viewers decide deserves treatment as a mid range resource user. at any rate it isn’t about what i want or think, i was hoping to just get people to think and debate it a bit more. and in fact it isn’t really about any particular setting or this particular setting that i think there needs to be more thinking and debate on. the real issue to me is, is it helpful, is it in the service of or on the side of caution to further divide people’s view of and experience of the same place and the same things? i think the popularity of invisi-prims (sorry i ever mentioned them now lol, but btw, they were not so great with semi trans or trans textures either) is insufficient reason to send the message that this setting and features dependent on it should be approached by everyone with caution. when i say everyone, what i mean is maybe there is a computer that on first install would get a recommendation of an ultra high graphics setting, but i have no idea what it would be comprised of, and i sure don’t run one. like you my 5 yr old clunker is a good quality one and gets me around in style, but it is hardly a high end machine at this point.

    speaking of which, curiously, in terms of caution a clean install of Firestorm starts me off with higher graphics settings recommendations and a higher LOD than a clean install of the LL viewer, which leads me to mistrust both equally in terms of what either can glean about my hardware capabilities or preferences are likely to be.

    @ Anseriel i had never noticed the semi trans and projector thing, i’ll go test it, but since you can’t get projectors except in deferred lighting, maybe the projectors are to blame somehow not the lighting mode(?). anyway i have used deferred lighting daily since it became available and also do have nice mesh feet with semi trans blending, so i will go shine a light on those puppies.

    Reply
    1. Ansariel Hiller

      Actually it is a common issue with how deferred rendering works. Semi-transparent objects also contain the color information of the object behind it and this will fool the lighting model, resulting in wrong colors. I also have a (NSFW ;)) video where the effect is pretty obvious as certain body parts suddenly turn into awfully wrong colors.

      Reply
    2. Inara Pey Post author

      Oh, I’m in full agreement as to trying to encourage discussion, etc. I’d rather see to it that people are informed as widely as possible through available means – blogs, forum posts, etc., both via LL and elsewhere sooner rather than later, even allowing for my earlier statement that not everyone reads blogs, etc.

      Leacing invisiprims / alphas aside (I also regret referring to the former in my reply), deferred does present a significant performance impact on the viewer. Until materials gain widespread use in SL, this is potentially going to leave a lot of people upset, regardless of the information available regarding the benefits for materials, etc. All people will see is a sudden drop in viewer performance between installs for little real benefit (as it is going to take time for materials to make their presence felt) oir real understanding as to why, despite any campaign to raise awareness. So there is going to be be shouting and gnashing of teeth and possible backlash. That’s got to initially weigh on any decisions made towards defaults, at least in the first instance, when it comes to changing things, particularly given the size of the FS user base.

      Which is not to say I necessarily agree with the FS team’s decision. Just that I can understand it – and as mentioned, it doesn’t have to always be that way. As materials usage spreads across the grid, there will be more time to better educate users (and for users to educate one another), and encourage them to try deferred and witness the results first hand. As such, I don’t perhaps see the Firestorm team’s decision as necessarily dividing people’s in-world experience to any great extent when it comes to seeing the impact of materials (as opposed to the impact of deferred on performance). And at the end of the day, it’s not as if enabling deferred requires the download of another viewer or a relog or anything – it’s just a tick in a check box in Preferences.

      Reply
  4. Rebecca

    I just wanted to thank whomever for the transcription. Also thank you to JL and EM and the whole team whose names I don’t know WF and others too. And the support group. Given I consider myself an intermediate-semi advanced level computer user and it made me sad only 2% read the blog. I see why you have the post up on the viewer log-in screen. I sure don’t know what else to say, but, I am glad LL is fixing the texture problem (I’m assuming this has to do with copybotting textures and their huge leak of info finally being closed.) so good for them! Huzzah! Hopefully they give everyone the time they need. Thank you so much everyone for doing this, unpaid. Thank you, thank you thank you.

    Reply
    1. Inara Pey Post author

      Thank you, Rebecca :). Putting the transcript together to a bit of time, but happy to have been able to do so.

      Gald you’re also enjoying your time with Firestorm, and I know the Firestorm team appreciate positive feedback, so thank you for that also. However, and without wishing to sound in any way rude, this blog isn’t directly associated with Firestorm, so I’m not sure as to how many of the team read articles / comments here. Might I suggest you consider posting the feedback also to their blog, so they will be sure to read it and appreciate it directly?

      As to testure problem – I assume you’re referring to the Server-side Baking – this may offer a sort-term solution for some elements of copybotting, but that’s not the primary aim of the project. It really is directed towards resolving the long-standing issue of avatar bake fail. However, as the new viewer side code might take a while to filter out into the wider world and malicious viewers, it may lead to something of a reduction in copybotting instances – but if so, it’ll only be unless / until those producing such viewers implement the code at their end. As such, I wouldn’t look upon it as a resolution to the broader issue of skin / clothing content ripping.

      Reply
        1. Inara Pey Post author

          If you look at the section under the video – below the LIKE THIS options, there is a series of page numbers (1 through 7) – click on these to read the transcript. Sorry it is not any clearer, but that’s how this particular WordPress template opts to place page numbers.

          Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s