Town Hall with Cory Linden, Dec. 20 at 2:30pm PST
Wednesday, December 6th, 2006 at 5:46 PM by: Jeska LindenWe hope you can all join us for December’s Town Hall with Cory Linden on December 20th at 2:30pm PST/SLT.This Town Hall will be a bit of a throw-back to earlier Town Halls, in that it will be text-only and carried via the official Linden Repeater. You can pick up the Repeater in the Linden Village in Kirkby.
Please bring your technical questions for Cory Linden as this will be a Second Life technology focused event. Have you always wanted to know what makes the grid tick? Wonder if we’ll ever finish Havok 2? Interested in the future of LSL? Bring your inner geek to the Town Hall!
What: Town Hall with Cory Linden
Topic: All of your technical questions
When: Wednesday, December 20th at 2:30pm PST/SLT.
Where: We’ll be gathered at the Pooley Stage, but please feel free to grab a Repeater and follow along anywhere in-world. (Or check the Event Calendar for listening parties!)
How to Ask Questions: You will be able to ask questions via the Linden Town Hall Questions group (you will need to join the group in advance)
Hope to see you there!
Please note, individual billing or support questions will not be accepted. Unfortunately, due to time constraints, it is not always possible for us to answer all questions during the Town Hall. Get there early.
December 6th, 2006 at 6:25 PM
Cory, I’d like to ask you a simple question: Why are you still supporting libsecondlife, and why are Lindens still in libsecondlife, when it has not reformed, when the people responsible for CopyBot, the god-mode stalking, and the jumbo prim griefing and destruction of the world all still remain in the group, as do W-hat fresh alts?
What is it you hope to accomplish by continuing your support of this group? What are they working on now that might be harming us next? And why is it that people in it are able to reverse-engineer the client, despite sanctions against reverse-engineering in the TOS?
Furthermore, I’d like to know what procedures are in place to protect Second Life from the griefing that occurs out of this group, and the machinations and supposed “accidents” that occur with untested, amateur-coded software that invades our world. Why do you permit software or scripted items of functions that can access SL that are submitted *anonymously* and posted in the library of libsecondlife or on third-party sites by libsl to be used within SL? And do you feel that Lindens should be supporting a group that gains financially from their work reverse-engineering the client?
We pay for a subscription to Second Life and we pay for server space and we sign a TOS; this TOS and this service doesn’t also say that we may be damaged or even have our businesses and second lives seriously harmed by experimenting on the reverse-engineering of the client by a series of griefing alts and anonymous hackers.
Please explain your continued support of this group also, in light of the founder’s admission that he has paid sources to lie to journalists, and also sold CopyBot on an alt.
December 6th, 2006 at 6:46 PM
Dear Cory,
I for one hope you DO NOT spend precious time talking about libsecondlife and falseAlarmBot.
The destruction of the world most definitely has not occured, the economy has not collapsed and the sky is not falling. With regards to support, it is a classic case of keeps your friends close and your enemies closer. These holes would be exposed sooner or later, better by those counted as friends.
The REAL issues are: Grid stability (network, database, server code) and Client stability (more graphic cards? no random crashes?)
How do you plan to support the massive growth which hopefully will only continue on its exponential curve?
Long live Second Life and long live hackers of all colors.
Peace.
December 6th, 2006 at 6:50 PM
The destruction of the world didn’t occur in part because the Lindens made unauthorized use of CopyBot a TOS offense, and the next version of SL they installed had elements that apparently “broke” some of CopyBot. But we need to hear about the extent of its potential damage — and the damage of future things like it — and the capacity of this reverse-engineering from Cory Linden himself, not from other residents who are speculating and trying to downplay the damage it has caused precisely because these hacker friends support and aid and abet the very people who are replicating and distributing griefer prims to overload sims *right now as we speak*. They are using the exact same M.O. as they have in the past, and some of them are in the group libsecondlife, and some key members of libsecondlife have supported W-Hat in the past. I would like Cory Linden to come to grips with this and explain *why*.
December 6th, 2006 at 6:52 PM
wow, judging by those two comments up top I’d say he won’t be short any content for questions.
Will you at least field this one:
Do you have anything to say to the people blaming free accounts for the lag and crashes and instability?
What sorts of things can be attributed to this recent influx of immigrants, and which ones are totally seperate?
December 6th, 2006 at 7:06 PM
Just an FYI, of _course_ the recent SL update broke CopyBot as well as every other client viewer that was installed. LL doesn’t like the fact that updates invalidate all clients and are hoping to change to a more flexible protocol, but in the meantime I was looking forward to this update “cleaning out the garbage” that was CopyBot. Of course, someone could update CopyBot to work with the updated client (which may take more work than simply commenting out a few lines of code), but at least this will drive it relatively underground so that everyone and his brother isn’t dragging one of those around…
December 6th, 2006 at 7:09 PM
One more question for Cory: How feasible would be for whole islands to be given access to the beta grid? Meaning, if I could collect a bunch of test rats to load up an island with 40 users, would that help on the load testing end of things?
@Prokofy: I won’t bother debating as we both obviously have set opinions. All I will say is if defending freedom of knowledge and speech is aiding and abetting then guilty as charged
December 6th, 2006 at 7:14 PM
Um, I defend freedom of knowledge and speech by making sure that tiny, elitist groups of unaccountable programmers don’t get to take these very freedoms away from *the rest of us* with their uncontrolled antics.
December 6th, 2006 at 7:15 PM
How very Republican of you
December 6th, 2006 at 7:23 PM
Well, in terms of LSL upgrades, there’s always the idea of C# that has been bantered about. Is it still on the table, and how would it work? What kind of includes would we have? Could we make our own “include” modules? A C# script would theoretically have as much access to the system as the includes would allow it to have (i.e. it probably would have the same “ll” functions that LSL has). Any idea of increasing the 16k thread limit?
December 6th, 2006 at 8:10 PM
Actually, I vote Democrat straight ticket except for when I vote Green Party, champ, go take your prejudices elsewhere, I’d like Lindens to answer what they know about libsecondlife; its past antics, its future antics, and their plans for oversight, or non oversight.
I’d like to hear their overall policy for open source projects within and around SL that involve them giving early heads-up, early access, beta access, “caps” or capacity for some, not others, etc. etc. I’d like to understand, when people sign on to this service, what sorts of things are going on under the hood that in fact are capable of disrupting the entire world.
December 6th, 2006 at 8:47 PM
I was one of the people pushing to hear from Cory - I hope the comments here can stick to the procedural and tech stuff … thanks Cory for doing this.
I would love to hear about:
* what are the current priorities for short term fixes, and what can you can change in the system architecture to handle.. say.. 5 million users?
* what is the deal with testing? (Kona and Brent are awesome, btw!) It seems like a big problem is not testing enough on a non-beta grid scale. What are you going to do to avoid more upgrade debacles?
* I hope that you post some sort of “outstanding issues and timeline” to the blog on a weekly basis - the community needs to see more technical communication and leadership.
once again, thanks for taking this on!
December 6th, 2006 at 8:50 PM
Well, I may not make it to the town hall, so hopefully my question(s) here may pop up:
1: What is the status of Havok 2 (3 or 4)? As an aircraft builder, I would really like to know when we can get more than 31 prims in a physics enabled object, this also reflects my wish for joints to be brought back.
2: What about the new rendering engine, any progress on that?
3: MONO to replace the LSL, any word on the development process and when we may expect to see a beta or other test release?
4: HTML rendered upon prims, this is one feature that many have been waiting to arrive for a while.
5: Misc items such as Object to Object communication instead of relying upon email routines and private chat.
Thanks,
Mike
December 6th, 2006 at 9:09 PM
All I’d like to know is: When do we get a more “modern” physics system?? Maybe support for PNG & GIF for texures like ‘other’ Virtual Communities have? Some have asked about being able to animate the Avatars fingers. Not sure how do-able that is, but something like this would also be of big interest.
December 6th, 2006 at 9:33 PM
Q: How is the avatar puppeteering comeing along?
December 6th, 2006 at 9:38 PM
A. Works great, Frans, we all log in like Punch and Judy every day and keep taking a beating!
December 6th, 2006 at 10:53 PM
Sounds fun, Corley. I’ll be there. Unless, of course, the Earth is destroyed to make way for a hyperspace bypass or the words of the prophets written on the sundry blogs occur.
December 6th, 2006 at 11:42 PM
Atm. most issues seem to stem from the limited scaling of both the central database and the central asset server. The short term solutions seem mostly to exist of quick fixes, and disabling functionality to prevent them overloading. Some longer term solutions have been implemented (e.g. baking textures on sims instead of asset server.)
Q. Are there any long term solutions planned to allow for seamless scaling of both asset management and database. And can you give any insight in what those will be?
December 6th, 2006 at 11:44 PM
Q: If packet sequence numbers are 16-bit values (unsigned shorts), why are appended ACKs 32-bit values?
Q: With the zerocoding compression used on packets, the only case where packet size increases is with a single zero value, which becomes two bytes (0×00 0×01) after compression. Since field sizes are already specified for string values in the protocol, wouldn’t it make sense to strip the null terminators on strings?
Q: I’ve been told that in profiling, Second Life spends almost 2/5 of it’s CPU time in the Kakadu library decoding JPEG2000 textures, and that while an older version of the library is being used in the client currently, there is a branch in testing that uses a newer version that takes advantage of multiple processors. Is there any time line on merging this in to the production code?
December 6th, 2006 at 11:44 PM
[...] Linden are holding another Town Hall on December 20th at 2.30 pm, which converts to December 21st at 9.30am Sydney time. [...]
December 6th, 2006 at 11:51 PM
Lol prok! =)
Prokofy++
December 7th, 2006 at 12:28 AM
I can’t be there, someone please ask these questions for all of us:
Cory, was the 1.13 update load tested in a staging environment before being deployed?
If not, will future updates be load tested in a staging environment before being deployed?
If not, why not and what is being done to prevent future updates from wreaking havoc like 1.13?
(By “load tested” I mean a scripted *simulation* of 15,000 concurrent logins.)
Do you have a formalized Quality Assurance process and team?
If not, why not and when are you planning on starting to recruit an experienced QA manager and team?
December 7th, 2006 at 12:31 AM
Prokovy: libsecondlife is a tool, not a griefer. A tool no different than those used to found, enhance, and yes, hack the internet today. All great projects start off as tinkertoys of the clever. Without these tools, the internet as you know it, your favorite tools, browsers, and protocols, would likely not exist. This sky-is-falling-mentality is absurd. To shift the blame for whatever problems CopyBot causes — which in fact is only one utility part of libsecondlife — is to completely deny personal responsibility in the use of said tools.
December 7th, 2006 at 12:39 AM
I would like to know how LL manage to push broken update after broken update out of the door.
December 7th, 2006 at 3:35 AM
The issues on everyone’s lips at the moment are…. wait for it….. two words…. “Grid Stability”, and a lot of people are getting really really fed up with the constant problems. There are people on the verge all over the grid, and it’s understandable. I can understand that maintaining such a network, database and server farm must be a *massive* task but it seems that the updates are not tested fully before being released. Then an update is issued and brings down the grid - or at least cripples it. I have heard people ranting about not wanting “pointless” updates on the clients, such as features, tweaks etc. They want a stable client, and a stable grid - and if they are paying for them, then really they should get a say. I understand client updates may be needed to resolve issues caused by those clients, but otherwise why bother issuing a client just for a few pointless additions?
So… my questions are this;
1. Is LL going to define a strict update/test/deploy policy?
2. What caused these database load issues?
3. What is being done to prevent a repeat performance?
4. Is more priority going to be given to stability/performance/scalability
Lastly, I’ve heard people say that paying members should get priority over non-paying ones - I don’t agree entirely (yes, I’m a paid one *smiles). Remember free members buy things too! Hence the reason LL not closing down new signups. I really feel if LL don’t start to get their act together, they are going to face some sort of back-lash and people just won’t bother anymore - which would be a shame.
Come on LL, get it sorted! Hope similar questions/concerns will be answered and not side-stepped!
December 7th, 2006 at 3:51 AM
Cory,
First, I welcome the return of a Town Hall in the “old” format, no matter what disadvantages it had. For me, two advantages will always be at the forefront of chat THs — the ability to multitask, and the ability to get a non-audio transcript that is easy to search for. Although I fully understand the so-called advantages of an audio chat — being able to cover much more during a limited time — text-chat town halls have other advantages as well, like the cool ability to be listening in with some friends to a Repeater and being able to make comments as we get the chat out of the cute Linden radio. With voice THs, half of the questions I see is mostly “what did he say?” or “did he just mentioned that the stipends were going to be cut for all users or just new ones?” etc. It’s nice to have one in the old format again
Anyway… to some questions now. I’m afraid people keep forgetting that the “copybot” is old news and I’m personally tired of the continuing hysteria which seems never to die. Instead, since this is going to be a *technical* Town Hall meeting, I would suggest that you’d take the opportunity of doing what Philip (and you) did early last year — let us know the milestones for the upcoming semester (or at least quarter) in terms of features.
Again, I’m not interested in knowing what bugs are going to be fixed or not — I know the goal is to fix *all* bugs in the queue. Eventually. Thus, addressing the issue is really just a question of “when”. “Eventually” is as good an answer as any other — all software development companies with a will to be around for the next years will obviously fix bugs
The only things that interest me is what is going on with the technology that is being developed, some of which is known to exist, but never deployed:
- the “physical avatars”, as Frans mentioned, which was shown to the Octocracy in June, so we know they exist; will they get also a better, higher-density polygon mesh?
which is usually the best I can squeeze out of a cellular card-based setup.
It’s rather clear that this will never see the light, but what additional things will be incorporated over time?
- the port of LSL to run over Mono, which was shown to work in December 2005 by Babbage Linden
- the change from UDP streams for communication into a RESTful, HTTP-based system, which we sort of know that has being in development for around 18 months or so. Getting always frustrated in demos done in public spaces (publics, libraries, colleges, some corporate environments) where the only open port is 80 — and that through a proxy! — I’m having increasing problems in doing successful demonstrations, and nobody is impressed with the unreliable wireless demos with 3-4 fps, high packet loss, and ping time of 4000 ms or more
- HTML-on-a-prim. It’s now one year since the uBrowser project was launched, over 20 months since in-world HTML was announced, and the best we saw so far was the Help function and now the login screen and the Web on the profile. I don’t see any difference, in spite of the new Mozilla libraries — input boxes, textareas, check/radio boxes still don’t work under the Mac version.
- Multiple clothes/skins layers that can be changed at will and dropped on your avatar as you like (allowing, for instance, skins to be developed in a multi-layered fashion, where skin proper, tatoos, and makeup to be on different and separately-controlled layers)
- What happened to the 2.0 Renderer? The last time anyone heard about it, well over a year ago, was that parts of it would be “slowly being deployed in the current renderer”. The first bit that was shown was on the Map; 1.6 or 1.7 million people never even saw the “old” Map (and this was just over a year ago!). I still have the image of “millions of prims in a sim” created by Ben Linden ages ago engraved in my memory
Then we have the “officially abandoned” projects:
- Port to Havok 2/3. Yes, I know
- The Jabber IM project. Sometimes this is flagged as abandoned, sometimes some Linden publicly announces that it’s still being worked on.
- The “lightweight” SL (the “mobile SL version” abandoned in 2004)
- The multithreaded client, where the viewport would be separate of the rest of the UI (I know that SL has *some* multithreading ability these days — but that the UI continues to be tightly embedded inside the codebase, instead of being something separate from the viewport, like what happens with things like OpenCroquet,There.com or IMVU)
- A model where asset servers were distributed in a completely different manner (ie. no centralised system, but a “federation” model where each sim server acts as an asset server, and UUIDs get “re-interpreted” to signify a pair [sim name;asset UUID], thus not needing to rely upon a centralised model any more). I understand that for some reason this approach was abandoned because it was way harder to track where some asset failed to get loaded than on a centralised server (where either the asset is there, or it has disappeared from the grid, thus acting as a “checkpoint”)
That besides all the cool features that we have no inkling that are being prepared and were never shown:
- support for multiple grids (LL-run or not); see also the last point on the “federation” model for the asset servers
- port of SL to Sun’s Project Darkstar
- the new Second Life Protocol
I guess that addressing all these points and establishing goals for meeting them, which could be announced, should be a good policy to implement. Both you and Philip did something similar early last year. And guess what — you even accomplished all you’ve promised in the timeframe (except for HTML, a notorious exception), even if some features announced for “late June” were actually introduced in September (”late summer”). But the important thing was that LL had precise goals to reach, from a features point of view, these were made public, and you endeavoured to reach those goals — and you did kept the promise! (Now, don’t some residents have short memories?
)
December 7th, 2006 at 4:13 AM
Damn I didn’t have my ad set to auto-update. It ran out on Tuesday and I had to create another one for L$50.
December 7th, 2006 at 4:21 AM
This post above is in the wrong blog entry. The first time I sent it I got a ‘page can’t be displayed’ message, then when I went to re-write it I clicked on the wrong entry. Oops.
December 7th, 2006 at 5:41 AM
No, Gwyn, Copybot is most definitely *not* old news. Because it’s not about some copying mechanism that may or may not still work. You and all the other tekkies were telling us all along that “this is the Internet and it copies,” so *that* didn’t go away, and that’s what it’s about. It’s also about leaving in the hands of arrogant, juvenile, inconsiderate young people the very features of our world. Is that on? I don’t appreciate it — and a lot of others don’t either. Since Cory Linden is a grown-up, I’d like to hear his plan for controlling these kids. That remains very relevant, given that these kids have produced 3 things already that have griefed and disrupted: CopyBot was only the latest; before that it was the jumbo prim, still for sale on Slexchange.com, and the god-mode stalking. We’ve even had conflicting statements from LL on whether the jumbo prim is legal or harms servers or not; and CopyBot, for all your pooh-poohing and dismissals, Gwyn, and claims that it is yesterday’s newspaper, is *a TOS offense*. Now, we have to ask — why would the Lindens make something that is outdated, broken, and yesterday’s newspaper a TOS offense? Are they daft? Let’s hear you talk your way out of *that* one.
December 7th, 2006 at 5:42 AM
Re: >Prokovy: libsecondlife is a tool, not a griefer. A tool no different than those used to found, enhance, and yes, hack the internet today. All great projects start off as tinkertoys of the clever. Without these tools, the internet as you know it, your favorite tools, browsers, and protocols, would likely not exist. This sky-is-falling-mentality is absurd. To shift the blame for whatever problems CopyBot causes — which in fact is only one utility part of libsecondlife — is to completely deny personal responsibility in the use of said tools.
No, libsecondlife is a tool made by a group containing many griefers that has repeatedly — at least 3 times — now been used for griefing and really serious disruption of the world and really serious loss of business income for the businesses on it. So yeah, have a sandbox with tinkertoys. Just don’t advertise it as a place where land as value and can keep value; where your IP is protected with a permissions system. If you do that, it’s misleading. I’d like to hear Cory Linden, not the copybotters themselves who have already proved their immorality numerous times, to comment on their suite of programs and utilities, and whether they are good or bad for SL *as we know it*.
Oh, I’m not shifting any blame — because the tools *were used by the members of libsecondlife itself to grief*. Don’t you remember Eddy Stryker’s alt Prim Revolution?
As for all those projects you’re asking for Gwyn, one can only say if they are putting them in the next patches on the schedule, then what other functionality can we expect to lose? How much sales will be lost? And go by real people’s real sales numbers they provide, not by the vapour statistics on the web.
Seriously, now that it is abundantly clear that every single patch going in now is a very deep political decision that is literally and very graphically pitting one group of consumers of SL against another consumers, what can be expected for the continuing losses of the groups?
If every time you put in something like hidding online status to reward macropayment lobbyists, SEARCH has to be closed and lose business for micropayment earners, then what are the prospects for small business on SL?
I would say the revolution is pretty close to eating its children at this point.
December 7th, 2006 at 5:56 AM
Talk to me about the current update process. Beta testing for patches before they hit the grid, when code bases are closed, the process for examining how a patch went.
What’s in the cards for Search functionality? Current testing methods, plans for the near and far future. Any interest in opening up read access to the information currently in search so you can get out of the search engine business?
Scaling. What are the current bottlenecks? How are they being addressed, and how will these changes affect our understanding of how SL works?
New functionality has been added to LSL in the last couple patches. What other features are planned for it in the next year? Are MONO and other scripting possibilities on hold or abandoned?
General tech and feature plans for the next three months and the next year. Anyone realistic knows they aren’t guarantees, but it’d be nice to have a clearer view.
December 7th, 2006 at 5:58 AM
What a cluster. Overall, it gets worse as time goes by. And the Lindens just keep running the bilge pump.
SL has more holes in it than swiss cheese. I wouldn’t get too comfy living here. The BlackHats are coming, I’m sure.
December 7th, 2006 at 6:17 AM
I have a simple technical question.
I am a content creator in Second Life, and I sell things to people. Two weeks ago, I sold some lindens and attempted to transfer the proceeds to my PayPal account. According to PayPal, the proceeds have not transferred yet, and it has been two weeks … Where did the money go? I can’t afford to lose money in any way shape or form!
Lynn Kukulcan
December 7th, 2006 at 6:51 AM
Gwyn,
it is an illusion that a team of 50, a fraction of which developers, can deliver all projects in the pipeline with an acceptable margin of quality given any timeframe as long as Google does not get there “first and faster”. Fast and Furious, which is the only option LL has, leaves a backlog of backfiring issues at all levels. Be afraid of delivering the show when the necessary side to it is to shutter the end users.
As we see each incremental update, the most recent of which a worthy one, leaves and disseminates a number of issues that stay unresolved for a while, from drought and/or partially submerged sims to search dead for weeks. QA is a real problem here. Support and help is available from all who cares, just don’t be afraid to ask.
Prokofy I’m sick of your obsession w/ CopyBot. LLs did what they were supposed to. I am aware as they probably are that the platform has to open up at some point, other projects of a similar venue are catching up quickly, what we’ll have in a year from now is the audience, not the show anymore.
I am not scared to see the search engine outsourced. But to keep it down for so long anticipating load issues when a replicating server can be set up in minutes, that is, with Proky’s argumentation, not recognizing that a “middle class” exists and is suffering real loss from search taken down for so long. I buy his take on this matter at least as a metaphor.
At this point in time and given the bandwidth required to run SL it is unrealistic to believe that we will see wider acceptance of said protocol within larger environments where bandwidth is critical. Already several projects are being aborted because of that aspect, where other collaboration-centered tools are intensively being investigated (OpenCroquet). Again we face an understaffed LL here, what optimization can ever happen on the Mac client, a quite literal and raw recompilation?
Dream on, thats for free, but keep your feet on the ground, we need both the show and the end users.
December 7th, 2006 at 7:16 AM
Hello, I had a few questions to ask:
First, Why is the restriction for prim sizing so small now? It would really cut down on prim usage AND lag if we were allowed to make larger prims.
Other qustion is, do you guys ever plan on making an option for prims that would let us cut down on the tries/polys? ive seen these prims in wireframe and there just loaded with tris/polys and i think it would be a neat option if we could lower the tris or maybe supplied with lower tri prims for basic stuff IE skyboxs, roads, simple building walls, floors, etc.. this would again help with the lag.
Thank you.
December 7th, 2006 at 7:29 AM
I’d like to know if we’re ever going to see SecondLife v2.0, or will we have to endure through v1.99.99.99.99.01b type of version-numbering scheme for a few more years?
Also, on the graphics side of things (maybe not your area of expertise, but something to think about), I have often noticed, when disabling camera constraints, I can pull my camera under the land and see water under everything. Looking up through the water creates some interesting fluid-motion distortions and reflections. This seems like a VERY processor intensive effect to be calculating everywhere on the grid - even in landlocked areas, since the water seems to be beneath the land everywhere.
Also, something else I have noticed, but haven’t heard any mention of it, so maybe I’m barking up the wrong tree here, but…
I’ve noticed some VERY HEAVY lag issues every time LL does something cute to the sun/moon. For example the Easter Egg sun during Easter, and the 1,000,000 sun during the one million mark celebration. Each time something like this is done we experience very noticable degredation of performance. Could it be that SL is its own worst enemy as far as lag and packet loss issues are concerned?
@Prok; Don’t let those zit-faced little script kiddies get you down - keep up the good fight. It’s actually kind of cute how they keep pulling your hair to get your attention like a little kindergarten love affair. You have to realize that they probably haven’t matured enough to recognize the feelings they’re experiencing in their little nether-regions, so they express their affection by pushing little girls in the playground or dipping their ponytails in inkwells.
December 7th, 2006 at 7:37 AM
“At this point in time and given the bandwidth required to run SL it is unrealistic to believe that we will see wider acceptance of said protocol within larger environments where bandwidth is critical.” - Starcomber
This is a good point. What, if anything, are you all doing, or planning to do, to to lessen the bandwidth requirements of the Second Life platform?
Also, I agree with Gwen that repeated text-based town halls are much preferable and am eager to see the answer to her question about storing uploaded mulit-keyed textures on sim-based asset servers.
December 7th, 2006 at 7:52 AM
[...] (more…) [...]
December 7th, 2006 at 7:59 AM
Thanks for making the complete list Gwyneth. I would like to see answers to all those questions as well.
And this remembers me to this, I met Gwyneth during the Doctorow book reader design contest(summer? 2005). My design was pretty succesfull and still I have people asking about it. But I decided not to make a product out of it, because HTML on a prim was comeing soon(tm).
Q:Cory should I spend time on finishing up the Bookreader or any other thing with tech like Xytext, textures and/or quicktime. Or is soon(tm) finally going to happen in say, the next 3 months?
December 7th, 2006 at 8:02 AM
Will grid performance ever improve to the point that SL was as usable as it was before 6/6/06?
December 7th, 2006 at 8:04 AM
I am enjoying this thread. There are a lot of good points being brought up.
However, as much as some of you might hate to hear it, Prokofy owns this thread to this point.
We can not proceed with this platform in this manner where introductions of disruptive programs and codes are being allowed into the world, that are greatly impacting its residents in a negative way, by a group being supported by LL.
LL has to get a grip of this situation. If their decision is to support such a group, then they should make a concious decision to immediately act on the negative repercussions that the group’s actions may have on the world, and make the situation right again. Not take a hands off approach, saying “there is nothing we will do about it because it will just happen again and we dont want to become involved in an arms race.” That is irresposible and unfair to the thousands of residents who depend on LL to protect their interests.
I am all with the “let’s press forward” progress movement into the future. But not at all costs. There has to be a balance. And that balance will be reached when every step forward is taken with the paying customer’s best interest at the forefront.
Proceed with caution.
December 7th, 2006 at 8:39 AM
starcomber Vig… I do not want search ‘outsourced’ and handled by an external company - or, even worse, multiple places.
You’ve seen the disaster that the forums became, with interested residents having to now keep up with a dozen forums instead of the one that covered everything.
Can you imagine the disaster of having to pay a dozen search sites just to get a listing on each of them in the vague hope that people looking actually look on the ones you’ve listed on?
It’s a very, very bad idea. Second Life should be a one stop, one size fits all, centralised place for everything - not handfuls of subcontracted sites all duplicating the same thing. Especially if, as is likely, one of the ‘pet development companies’ ends up getting rich whilst everyone else suffers.
What guarantees do we have that any third party site will be impartial? This store is run by a friend, so I’ll tweak it so they get premium placement regardless of whether it fits what the person searched for. I don’t like that person, I’ll tweak it so their store only appears one in twenty hits. You don’t think that will happen? Trust me, it will.
Broccoli
December 7th, 2006 at 8:45 AM
Gwyn, all the things you’ve listed have already been asked at previous townhalls, and you already got the answer, “Coming soon…or when we get to it…next!”
That’s why I’d like to hear a more zoomed-out meta take on what the Lindens’ plan is in general for this entire almost-organic wild code garden they can’t seem to tame. Are they still busy swapping the engine out of the plane at 30,000 feet? Could they start Second Life 2.0 over on some other server farm and move some element of SL out of it? Will they license the software to businesses even if they don’t open-source it?
“Prokofy I’m sick of your obsession w/ CopyBot. LLs did what they were supposed to. I am aware as they probably are that the platform has to open up at some point, other projects of a similar venue are catching up quickly, what we’ll have in a year from now is the audience, not the show anymore.”
It’s not some literal obsession about “CopyBot” — which we’re told is broken or passe or harmless (depending on who you talk to). It’s about the larger issue that CopyBot signifies — out-of-control hacker groups larded with really serious griefers where Lindens are in the group. What is Linden Lab’s plan for overseeing and restraining this group *they* have responsibility for by a) having Cory glorify it here on the Blog and b) having Lindens in it?
>I am not scared to see the search engine outsourced. But to keep it down for so long anticipating load issues when a replicating server can be set up in minutes, that is, with Proky’s argumentation, not recognizing that a “middle class” exists and is suffering real loss from search taken down for so long. I buy his take on this matter at least as a metaphor.
I want to hear TO WHOM they have outsourced this job, or whether they have signed NDAs to have some groups get a head start on working on this, or what their overall plan is for SEARCH.
Ultimately, I want the technical people who keep brushing away the social concerns and insist on talking tech talk to explain and justify their *social and political decisions that THEY make*.
December 7th, 2006 at 8:47 AM
I got a quiestion. why should we pay for this game if we keep getting bs from ll and live help just sucks
December 7th, 2006 at 8:51 AM
“We pay for a subscription to Second Life and we pay for server space and we sign a TOS; this TOS and this service doesn’t also say that we may be damaged or even have our businesses and second lives seriously harmed by experimenting on the reverse-engineering of the client by a series of griefing alts and anonymous hackers…
I’d like to hear their overall policy for open source projects within and around SL that involve them giving early heads-up, early access, beta access, “caps” or capacity for some, not others, etc. etc. I’d like to understand, when people sign on to this service, what sorts of things are going on under the hood that in fact are capable of disrupting the entire world…
So yeah, have a sandbox with tinkertoys. Just don’t advertise it as a place where land as value and can keep value; where your IP is protected with a permissions system.” - Prokofy
These are reasonable observations and the requests that deserve solid public answers. In fact, now that the question has been clearly posed, Cory can’t not answer. Avoidance or ambiguous crowdspeak will be a clear answer that SL *is* nothing more than a hacker sandbox to LL meaning that land and IP have no value. We can then all decide whether we want to continue participating in the project.
December 7th, 2006 at 8:58 AM
… and i’m quite sure that those able to attend won’t let him fluff over these very serious questions too.
Broccoli
December 7th, 2006 at 9:24 AM
[...] Awesome comment by Gwyneth Llewelyn in Linden Lab’s official blog post on the upcoming Town Hall with CTO Cory Linden. [...]
December 7th, 2006 at 9:40 AM
Broccoli, Prokofy, outsourcing services is not taboo.
While I am not especially pushy about search being outsourced I re-render here that if it was the world would not, definitely, come to an end. Liabilities and ethical issues are there of course, potential or, with Prokofy, more real than real. But you are dreaming of a level of control and centralization that is not even possible in technical terms nor convenient in economical terms. The future of SL is not more control over everything, please see my statements under this, inevitable, truth.
December 7th, 2006 at 9:52 AM
If the future of SL is that they are rendering land and IP null and void, we need to hear that now, not be continually suckered into it.
If they just wish to produce a software, God bless them, produce away — outsource, insource, it’s your game, do what you want. Just don’t imply that everyone coming to it has a stable, productive *world*.
And there are certainly ethical issues in handing off the SEARCH work to one of their special insider companies, or equally, handing it off to some impersonal large company somewhere. The search was working last week. We can’t understand why it can’t work this week and there is a *political, not technical decision* to shut it off.
If the idea is that they need to have 2 million accounts, 75 percent of them free, to create the illusion of masses for big business to come in, that’s fine, then they have to say, sorry, you 10-25 percent other people who were trying to serve these masses with creations and land development. You’ll have to go away now, we’re done with you.
December 7th, 2006 at 9:58 AM
If we get html on a prim does that mean we can get an image from a website on a prim? If so, how will you prevent that SL becomes a land full with “error 404/401 etc etc” textures when people abandon their builds for a few weeks/months (Although… it would be kewl with the recent “image not found” avatar..).
Read and type you later
aEo
December 7th, 2006 at 10:17 AM
I too hope we can skip the whole copybot/LSL debate. Everybody’s made their positions clear- LL isn’t going to get into an arms war to stop the unstoppable, and some content people will continue to be dissatisfied as long as their creations are not ‘protected’. LSL isn’t considered a threat, and those in charge support the concepts of open source. Content people still hate LSL and the ‘god moding griefers’.
At this point- IMHO there is very little further good to come of the issue. Both sides have explained their points to death. Copybot came, copybot went, the sky didn’t fall. Move on.
I’m mostly interested in hearing some tech info about what happened this week (for curiousity) as well as insight into upcoming features like html-on-a-prim
December 7th, 2006 at 10:19 AM
er wrong button
html on a prim, and if/how/when the focus beta changes might integrate into the main viewer…
December 7th, 2006 at 10:34 AM
I too hope we can skip the whole copybot/LSL debate. Everybody’s made their positions clear- LL isn’t going to get into an arms war to stop the unstoppable, and some content people will continue to be dissatisfied as long as their creations are not ‘protected’. LSL isn’t considered a threat, and those in charge support the concepts of open source. Content people still hate LSL and the ‘god moding griefers’.
Once again, we’d like to hear *Lindens* not speculators and boosters and cheerleaders on this subject. If the Lindens didn’t need to get into an arms war, why did they make “unauthorized use” of CopyBot a TOS violation???
And it’s not about this or that project Libsecondlife is working on. It’s about the whole concept. Example, here’s what John Hurliman has posted today on my blog.
Prokofy Neva: The Lindens are more likely to release whatever sniffing and scraping functions they need to release just to get search working BETTER. If that means handing it to a third-party they trust, one of their own groomed start-up companies, they will.”
John Hurliman: The keys have already been handed over so to speak. There’s still time to contact me in-world (Eddy Stryker), or other libsecondlife developers regarding search engine systems. The bidding process definitely isn’t over yet.
and: “And due to the competitive nature of the project we’ll need to sign an NDA first, so please contact me in-world if you are interested in the project.”
So I’d like to hear from Cory Linden why the Lindens are supporting an ostensibly open-source library project reverse-engineering the client, which they are supervising, that enables some people to create proprietary, for-profit search engines for use with SL.
Is this the direction LL is taking since they have shut off the inworld SEARCH PLACES?
December 7th, 2006 at 10:45 AM
What I’d like to see is something like a Google TechTalk on SL’s database issues. BTW, the Google TechTalk was great, and the Lang.NET Mono talk was awesome, I think you guys blew some people away with your Mono hacks. Don’t even mention Havok 2,3 or 4 in the blog unless you’re serious… which I don’t think you are.
BUT the big question is… if SL has problems with 10,000-20,000 concurrent users, will SL be sustainable with 100,000 concurrent, or *gasp* 1,000,000 concurrent users? My current hypothesis is that no, SL won’t work with the current architecture.
LL is no Google. People in-game probably create and modify objects about a thousand times faster than programmers upload webpages to the web. Like you said at Lang.NET, over 60% of people are creating content in-game, whereas 3% create things on the web. Not only that, GoogleBot indexes everything, but GoogleBot also “un-indexes” things when they are no longer available. With content-creation in SL there is no “un-indexing”… everything stays on the Asset Server forever, and nothing is ever deleted (as I understand it). I’d like to see a discussion on alternative methods on how to handle the rate of content-creation, given that every instance of an object can be thrown back on top of the massive array of database servers you have, even if only a small change occurs.
I think one way would be to logically separate the “content creation” process, and the “content use” in the database with a huge architecture change… but would it be feasible? If you are this far along with the current system, should you even try? I don’t know, but it makes for an interesting discussion… that’s why you should have some sort of TechTalk on it.
December 7th, 2006 at 10:49 AM
I still think LL should have stuck with the TOS and not given permission to libsl to do what they want and I think all libsl members should be banned from sl along with any group remotely interested in hacking/reverse enginerring or griefing. Sure they’ll create alts, maybe take the reverse engineering underground, sure anything they do will be “easy” for them but then again who cares, if they do it again, send the fbi after them and just because you can do something, doesnt mean you SHOULD do it. If you have any problems with the last part, look who happened to fly a couple of planes into the world trade center.
When I joined SL it was a far different place, pleople only had 1 alt, they were required to verify themselves before joining and the database for the most part coped with the load.
LL should have understood what removing account verification would do to both the environment and their systems and yet they moved ahead and seemed really surprised when griefing increase and their systems became virtually useless.
Outsourcing search is a really stupid idea. The problems with the current system are that it can’t cope with the load, and it can easily be fixed by moving the search queries to a seperate database and using a transaction database to store updates similar to what thousands of large corporations do all the time when they have to deal with so many queries.
As for staffing, it’s not the real problem as far as I can see. It’s managing the resources they have already ie - depending on customers to stress test the test grid. The upgrades should be stress tested even before they reach the test grid and it’s not very hard to create automated systems to do this yet after 6 months it’s still not being done.
Second life is not leading edge, it’s not new, and we shouldnt be expected to alpha test every update, I won’t even say beta here because most of the updates are nowhere near what a normal company would call release candidates.
LL might want to be a company managing a “platform” rather than a world but they didnt/havent/won’t create a platform until they can get their act together, until they create stable ways to introduce new features, and until they take a good look at their systems, stop patting themselves on the back and realise their own limitations.
1.6 million accounts might look great on paper but there’s really no point if it makes SL useless to the people using it and if LL can’t afford to update their systems then they should have considered making “premium” accounts much more attractive prior to opening membership. It seems quite strange to me that the huge numbers are attractive to corporate clients but seemingly completely ignored by LL.
December 7th, 2006 at 10:52 AM
Cory, please consider the following suggestion.
Although I like the town-hall meetings and the interactivity they provide, they are usually a chaotic mess. A lot of really interesting questions have already been posed here, in the comments to the blog entry. Would you consider pulling these questions out into a document, and typing up answers to them, almost as though you were being interviewed? I would prefer your thoughtful and thorough answers to these questions to the one or two sentence fragments you are able to type in a live chat format.
It’s just an idea.
December 7th, 2006 at 10:56 AM
Persephone - great suggestion! In the past we have asked questions gathered beforehand (usually in the forums) and I’ll gather them together and see if Cory could answer some of the questions here as part of his introduction to the Town Hall.
December 7th, 2006 at 11:01 AM
Prokofy,
I don’t think you are fooling anyone here. Everyone knows you have personal vendetta against a couple of the people involved with LibSL. Your attempt to pass this off as a technical concern is weak.
December 7th, 2006 at 11:01 AM
That is wonderful Jeska. Thank you, and thank you to Cory for making himself available for a more technical town-hall. I look forward to his thoughts and comments.
December 7th, 2006 at 11:03 AM
“I still think LL should have stuck with the TOS and not given permission to libsl ” -Harvey
The TOS has changed to specifically allow groups like LibSL. Reverse engineering has nothing to do with griefing. Don’t believe Prokofy’s FUD, he has ulterior motives.
December 7th, 2006 at 11:09 AM
First, this has been beaten to death so this will probably be my last post on the subject. It’s not meant as offtopic, just a response to what I see as a useful and slightly relevant discussion (what is discussed @ next TH).
IMHO the lindens HAVE spoken. Dig up the last town hall where Philip addresses it, it’s within the first 10min or so as I recall. Also dig up on google video the presentation he and Cory did for a handful of Google engineers early this year.
In the TH, Philip explains that the server has to send certain things (geometry/textures) to the client so they can be rendered on the screen, so it will always be possible to scrape up this data either thru network sniffing, hacked OpenGL driver, even if not a 3rd party client. Philip also clarifies that LL still supports open source and still values the presence of LibSL, who (personal comment) IMHO have learned their lesson.
In the Google video, about 2/3 of the way in as i recall one of the Google engineers asks how they address security issues like the godmode client. Philip’s response (as I recall) was that it did cause a privacy issue which they would look into, but mostly they saw it as feedback of what features people wanted, so they talked to a few godmode client users and learned which of the godmode features were most wanted, some of which were integrated into the main client.
Now I am speculating as I am not a Linden, but I would assume they made unauthorized use of copybot explicitly a violation because it was already an implied violation (as i recall). I think the old TOS had something about IP violations, not sure of the specifics. Now they explicitly say that using a hacked client to rip stuff off is a TOS violation.
Also note the specifics- using a hacked client TO RIP STUFF OFF is a TOS violation. Which means copybot itself isn’t illegal, but using it to steal things is. A very important distinction.
As for the rest of your quote- there isn’t much context there so I can’t really comment. However if it means that LL is hiring outside help to get another handful of brains working on search issues, great! If LL is bidding out the project to independent developers, so what? IMHO what that quote is actually talking about cannot be determined without much more context. Either way, the LibSL guys understand SL better than most other random people, so they would probably be best in a position to help with Search.
Also- LL is NOT (info coming from both LibSL webpages and Linden posts) ’supervising’ LibSL. LibSL exists with LL’s blessing and they talk to each other on IRC working together to squash bugs. But LL is not in charge of or supervising LibSL.
And if LSL or any other group is making their own search engine, great! Competition is always a good thing. If you are worried that LL shut off search for good and the nice easy Search button is going to be replaced with a hodgepodge of independent search providers, I highly doubt this will happen anytime soon. Maybe if/as/when SL becomes ‘web 4.0′ or whatever, and everything is open source and everybody runs their own SL server, but not anytime soon (thats personal speculation). It doesn’t make sense.
LL (info coming from this blog) shut off the search due to technical problems, not policy decisions. They have said many times it will be back ASAP.
Either way I expect this TH to shed some light on the database/search issues, so thsi will shortly become a moot point.
December 7th, 2006 at 11:10 AM
What I said has nothing to do with prokofy, reverse engineering WAS against TOS, I think it should still be with no exceptions, LL might not be in the job of being police but then thats what the FBI is for. It’s my opinion and if you don’t like it, tough.
December 7th, 2006 at 11:12 AM
As for more useful stuff:
Pers has a good suggestion, but to add to it- Just working off a list might edge out even more of the hundreds of questions asked and not gotten to during the TH…
I think that ideally a slashdot-style question system could be most helpful- people could submit questions before or during the actual event. People could also then vote for questions (using either a webpage or scripted object), and the highest rated ones would get fed to Cory / Philip. Before-show voting would mainly be to screen out a lot of the preshow questions and resident voting during the show might count for more.
This whole process would of course need moderation or hand-sorting afterward, ie at last town hall I could see 50 different wordings of ‘we don’t like copybot, kill it’ being modded all the way up….
December 7th, 2006 at 11:26 AM
One of the limitations, I understand, is available bandwidth. We can protect people’s IP rights, reduce load on the asset servers, and completely disrupt copybot development, all by shifting the responsibilities of each of these things around.
For example, we can have each sim have it’s own database of “rezzed prims.” This database would be what you get the sim’s world from - not from a central asset server - thus reducing the load *on* the asset server.
If we move the creation system to the server, then we again improve things, as well. The assets won’t go straight to the asset server - instead, they will go to the sim where you’re creating things. Also, the sim can check if you try to copy something if you in fact have permission to copy it before allowing such copying to occur. Which makes bulk copying more difficult.
It reduces bandwidth. Instead of “Create prim type with prim dimensions, textures, and yadda yadda yadda …” It sez “Create new prim of this type at this location” or “Copy prim X at this location.” Modifying would be similar. “Change prim X size to …”
Which elimates creating giant prims, too. Because … the server checks the max size, not the client.
It may not save a lot of bandwidth, but a savings of 1% on the bandwidth is, for the huge amount of bandwidth that SL obviously uses, still a significant savings.
We do need bigger prims. I would suggest that 32m x 32m x 32m be the maximum prim size. LL claims that they can not be sure that prims do not overhang other people’s parcels but … if the sims were handling the prims, they could do this themselves. This would allow the landholder to return any prims or objects that overhang their land as well as any prims or objects that are directly on their land.
A lot of the CopyBot supporters would, of course, be rather unhappy with the notion of checking prim overhang. I tend to disagree, what with llVolumeDetect and stuff already working.
I will grant my suggestions move the load from {a} the Asset Server to the Region Server, {b} from the Client to the Region Server, and {c} just add more load to the Region Server.
That being said, a lot of these suggestions would improve bandwidth and asset server load, and I think they would go a loooooong way to restoring system stability.
December 7th, 2006 at 11:37 AM
During the last town hall meeting, I was curious what Linden Lab’s plans for going fully opensource with the client were. Of course we got the standard, “We have no information on that at this time.” but I know it’s been brought up. Just what guarantee do we have that if SL goes opensource it won’t become an ‘elitist programmer’ paradise where if you don’t know at least 3 programming languages you might as well sit in the corner while everyone else has the fun? Part of what I like about SL is that with LSL scripting and prim construction, it doesn’t matter if you know Maya, 3D Studio Max, C++ or any other programming language. It’s all original and has to be learned by everyone when they join. This keeps things even and fair for everyone who joins in, allowing anyone, regardless of prior training to have fun and be successful. It’s part of the reason I DREAD the thought of allowing third party meshes to be imported. Sure, everyone who knows how to do it and can afford a $500 modeling program can do more, but it will basicly render the entire Prim system of building obsolite, making it so once again, only people who can afford high end programs or years of schooling have any chance of doing anything remotely successful.
More or less, since this is going to be a ‘technical’ discussion, what can you offer those of us without 6 years of hardcore computer training in regards to making sure that SL doesn’t become yet another elitist playground of experienced hackers and coders, and anyone who is just there to enjoy themselves is greeted with ‘Linux or GTFO!’? That is my main concern. It’s obvious that as smart as programmers are, for example with LibSL, their judgment on the human side of things can be somewhat lacking. I know the ‘Freedom of speech and technology’ is the battle cry for the modern programmer, but honestly, without secrets, laws, and guidelines there would be unending chaos because the nature of the human species is to create AND destroy. But I’m not going to whine over Copybot. It was an experiment gone wrong and I’m not out to linch LibSL over it. I’m sure they’ve probably done countless other things that have enhanced the quality of SL that no one’s ever heard about inbetween their lapses of judgment. But there needs to be a balance. SL is an network of code, just like a city is a network of metal and concrete. But you shouldn’t have to be a certified contractor to live in a city….
December 7th, 2006 at 11:43 AM
My 2cents:
Also, on the graphics side of things (maybe not your area of expertise, but something to think about), I have often noticed, when disabling camera constraints, I can pull my camera under the land and see water under everything. Looking up through the water creates some interesting fluid-motion distortions and reflections. This seems like a VERY processor intensive effect to be calculating everywhere on the grid - even in landlocked areas, since the water seems to be beneath the land everywhere.
This effect is only calculated when you (or someone else’s) camera is under water. This makes it a non-issue when it comes to framerate and the fact that water is everywhere.
Just an FYI…
December 7th, 2006 at 11:54 AM
I agree with Sasha Wheeler, too.
Lynn Kukulcan
December 7th, 2006 at 12:03 PM
I have no personal beef against anyone involved with the Copybot project - I, like many others, just feel that their current track record of:
- Huge Prims
- God Mode Hack
- Copybot
have proven, beyond doubt, that such ‘reverse engineering’ - whilst still a ToS bannable offense - is dangerous and nobody should be permitted to hack the client for any reason - let alone have Linden Lab support for it.
It is well known that a small number of people have made themselves a fairly decent amount of cash by the sale of these ‘tools’. Is there a reason why there has been no attempt to recover this money from the guilty parties?
Broccoli