Chat log from the meeting on 2017-01-31

From OpenSimulator

Jump to: navigation, search

[11:02] Nebs Metal Bar Stool v1.6 (w/sit & launch): Hello Andrew Hellershanks, enjoy your sit..
[11:03] Andrew.Hellershanks I'll be right with you. I have one cron job to update.
[11:07] Andrew.Hellershanks ok, done.
[11:09] Mike.Lorrey Whats the haps?
[11:10] Mike.Lorrey Hopefully I can stay online this week. Comcast cut out in the middle of last weeks meeting
[11:10] Sheera Khan: we have a growing number of users in our grid where the internet connection runs on IPv6... OS doesn't seem to handle that gracefully :-(
[11:11] Marcus Llewellyn: It doesn't much matter unless the viewer also supports ipv6, and that is unlikely to ever happen.
[11:11] Sheera Khan: so no fun running the simulator on the home-PC anymore for them ...
[11:11] Sheera Khan: it's a viewer side problem?
[11:12] Mike.Lorrey in my case it was a outage covering my whole town, so not an OS issue
[11:12] Marcus Llewellyn: OpenSim could, with some probably nontrivial work, support ipv6. But the LL protocol doesn't support it, and the viewer doesn't support it.
[11:13] Sheera Khan: hmmm...
[11:13] Sheera Khan: but the same people can connect with their viewers to IPv4 simulators
[11:13] Sheera Khan: even though the connection is made over an IPv6 link
[11:14] Marcus Llewellyn: Most ISPs still support ipv4, and probably will for quite some time.
[11:14] Sheera Khan: it is changing rapidly here in germany
[11:15] Sheera Khan: some don't even give you a dual stack ...
[11:15] Marcus Llewellyn: That may be, but ipv4 is too entrenched in too many apps to go away very quickly. If nothing else, ISPs will support things like ipv4 traffic encapsulated in ipv6.
[11:17] Marcus Llewellyn: And in any case, there's not much OpenSimulator can do without a ipv6 compatible protocvol stack and client. :(
[11:17] Misterblue Waves: there are a lot of IPv4 assumptions in OpenSimulator... there is a bunch of work to get it to work for IPv6
[11:18] Sheera Khan: :-(
[11:18] Mike.Lorrey One thing I am curious about is if its accepted to issue patches for older version of OS?
[11:18] Andrew.Hellershanks That is something that could be looked at once 0.9 has been released.
[11:19] Marcus Llewellyn: Mike: There tend to be "postfix" releases for a version ot two older than the current release, but it usually only includes important patches... security, or thing that might break compatibility (say changes to how TPs work).
[11:20] James.atLLOUD Sheera, what happens, do the viewers simply not connect, are there performance symptoms?
[11:20] Sheera Khan: @ James: they can't connect to the server
[11:21] Sheera Khan: so the simulator is running but nobody can visit them
[11:21] Andrew.Hellershanks Mike, patches might be accepted for 0.8.2 if they fix a bug. If it isn't an import bug fix I would suggest making patches for the current (git master) branch of code.
[11:21] Mike.Lorrey I am primarily interested in sim physics. I think the most important bug is the malassignment of keys in the ODE engine vehicle linear motor parameters
[11:21] James.atLLOUD OK, ty Sheera.
[11:22] Marcus Llewellyn: I'm gonna presume that for people hosting from home, the ipv6 breakage is prolly due to being behind a NAT.
[11:22] Mike.Lorrey while a lot of work has been done on ubode, a lot of people are going to remain with older versions of OS so unless ubode offers backward compatible updates for them, a patch for the ODE, which I've seen in operation that is very simple, should be seriously considered
[11:23] Mike.Lorrey Carious at Awesim Grid has the patch for ODE in operation on his grid, he has every sort of physical vehicle operating in ODE
[11:24] Sheera Khan: @Marcus: you could be very right... i suspect most home networks to run on RFC1918 addresses still ( et al i.e.)
[11:24] Andrew.Hellershanks There isn't anyone actively maintaining ODE that I know about. Most work for the physics engines is on BulletSim and ubODE.
[11:25] Selby.Evans Kitely might use ubODE, if it would run on 8.2
[11:25] Mike.Lorrey I am aware of that issue. Who can I talk to about putting together a group to take over ODE if nobody else is handling it?
[11:26] Marcus Llewellyn: ODE is dead, except for Ubit's very fine work. ODE's problem is that as a project it is more or less morbund. It sees very little active devlopment these days. It's more sensible to move on to an engine that is actively developed.
[11:26] Andrew.Hellershanks Mike, why do you say that a lot of people are going to remain on older versions of OpenSim?
[11:26] Mike.Lorrey Because I see today a lot of people esp on OSG, who aren't even operating 0.8.2, so it stands to reason they won't be updating to 0.9 either.
[11:27] Misterblue Waves: the changes that ubit have done on ubODE is in the C# code... ubODE and ODE the same ODE binary
[11:27] Sheera Khan: as much as I'm for choice... but it makes scripting more complicated and error prone to have several engines to script for^^
[11:27] Mike.Lorrey if it is the same binary, why isn't it able to be installed on older versions?
[11:28] Selby.Evans Ilan did not state his objections to .9, but he seemed adamant
[11:28] Marcus Llewellyn: OpenSim ODE physics is not the same as the binary. The binary enables OpenSim physics. It doesn't provide the nitty gritty implementation details we all experience inworld.
[11:28] Misterblue Waves: if there are some showstoppers, we'd like to hear about them
[11:29] Sheera Khan: @Mike: you'ld need to backport Ubits changes to the C#-code to your older version ...
[11:30] Mike.Lorrey and Ubit isn't doing this himself why?
[11:30] Andrew.Hellershanks If some don't update their regions/grids to at least 0.8(.2?) they may eventually find they run in to issues when attempt HG TPs in/out of the region/grid.
[11:30] Marcus Llewellyn: And even then you'd prolly see some script breakage, and UbODE behaves closer to SL's physics. Classic ODE had very different behavior.
[11:30] Sheera Khan: @ Mike: maybe he is busy getting 0.9 forward ^^
[11:31] Misterblue Waves: even the viewers will stop talking to older simulators.... a bunch of the control channels are moving to TCP
[11:32] Andrew.Hellershanks Ubit is focusing on the current version of code. 0.8.2 is a stable release. A stable release will only receive bug fixes.
[11:32] Sheera Khan: with the limited manpower available OS might not be able to support all legacy simulator releases ...
[11:32] Marcus Llewellyn: In my experience, the people clinging to classic ODE do so because they want their script that uses physics to keep working just like they always did. It's a painful but necessary realization to make that that blocks a better and improved physics experience.
[11:34] Simulator Version v0.5 ruft: OpenSim Dev 056ccae: 2017-01-27 21:43:25 +0000 (Unix/Mono)
[11:35] Mike.Lorrey I don't have any problem adapting scripts, that's what i do. What I want is for people who are supposedly running this train to consider fixing the wheels on the whole train, not just the ones on the current shiny car
[11:35] Andrew.Hellershanks From what I have heard, ubODE has better compatability with SL physics. That should make it easier to use physics related scripts that came from SL. If a script was written for, or adapted specifically for, OpenSim then it would depend on what script engine was in use at the time and what changes may have been made to that engine since the script was written.
[11:35] Misterblue Waves: some of the vehicle parameters in ODE are way off when compared to SL.... both ubODE and BulletSim try to be closer to SL behavior (script parameters to what happens)
[11:36] Kayaker.Magic Mike is talking about some ODE parameters that HAD THE WRONG DATA TYPE!
[11:36] Mike.Lorrey ODE parameters are messed, as i said, because the person who coded it couldn't reconcile the linear motor acceleration with buoyancy and gravity and simply gave up. That oversight has been fixed and there is a patch available. Why not use it?
[11:37] Andrew.Hellershanks If you keep back porting some of the latest changes/improvements to the older versions of code it would slow down development on the latest version of code and also reduce some of the reasons for people to update to the latest release of code.
[11:37] Marcus Llewellyn: It simply isn't realistic to expect the core devs to maintain old versions, with the exception of postfix releases. If grid owners want to update old versions, the onus falls on them.
[11:37] Misterblue Waves: Mike... is the patch on a Mantis entry?
[11:37] Mike.Lorrey that sounds like the sort of excuse that Microsoft would make to force people to buy their latest version of crap. There isn't any profit motive to justify such excuses.
[11:38] Mike.Lorrey I do not know if Carious has posted his patch to the mantis. I will bug him about it.
[11:38] Misterblue Waves: but then there is the problem of becoming incompatible with scripts that were tuned to work with the legacy ODE
[11:38] Mike.Lorrey From what I've seen, the fix doesn't make scripts tuned for ODE to not work
[11:38] Andrew.Hellershanks Mike, if you are a scripter then you should understand how code is handled in various projects.
[11:39] Mike.Lorrey because those "tunes" are really just kludges using bouyancy as an accelerator
[11:39] Misterblue Waves: I guess there could be a region configuration parameter
[11:39] Andrew.Hellershanks You have a stable release and the latest development version. The last release gets bug fixes only and the development version gets all the new stuff.
[11:40] Mike.Lorrey if a nonperforming physics engine isn't a bug, I don't know what is
[11:40] Andrew.Hellershanks If there is a patch that could update ubODE in 0.8.2 (or earlier) then a grid owner can apply it if they really want the recent changes in ubODE without updating the rest of the grid software.
[11:41] Marcus Llewellyn: I've personally been under the impression that classic ODE is considered deprecated. Whether it is or not could perhaps be more clearly stated.
[11:41] Andrew.Hellershanks It isn't a configuration that would be supported by the core developers.
[11:41] Mike.Lorrey lol
[11:41] Mike.Lorrey of course not
[11:42] Andrew.Hellershanks Classic ODE is pretty much dead.
[11:42] Mike.Lorrey if the internet was developed like the core developers mess with OS, we would never have gotten past HTML3.0
[11:43] Sheera Khan: @ Mike: since OS doesn't cost the user anything your argument can be used in reverse ^^
[11:43] Mike.Lorrey that it is worth exactly what you pay for it?
[11:44] Sheera Khan: maybe that you have to take what you get for free ^^
[11:44] Mike.Lorrey as I recall HTML and apache has always been free too
[11:44] Marcus Llewellyn: Or how much effort *you* want to put into it. Expecting core devs to bend to everyone's whim is not a good way to look at things. Especially if that whim is maintaining deprecated code.
[11:44] Sheera Khan: or do it yourself...
[11:44] Andrew.Hellershanks Mike, using your argument you would just backport some of the improvements and/or new features of HTML 3.0 to HTML 2.0 and a lot of people would stay on 2.0
[11:45] Sheera Khan: and how many servers run apache 1.0 anymore?
[11:45] Sheera Khan: or websites using HTML 3.2 standard?
[11:47] Sheera Khan: and OS is not even at 1.0 ^^
[11:47] Mike.Lorrey opensim is already fragmented into many different forks because of the political intransigence of a small clique, because people refuse to work together on fixing bugs for everyone and working to international standards.
[11:47] Misterblue Waves: people need what they need, though..... if it's doable and don't mess too many people up, it should happen
[11:48] Mike.Lorrey where is the international opensim standard?
[11:48] Marcus Llewellyn: Prolly on the shelf next to the Linux standard. ;)
[11:48] Mike.Lorrey why won't IEEE or other standards body even consider it?
[11:49] Sheera Khan: a creator af a vehicle only needs to add an additional branch for a different set of parameters to make his creations run on modified ODE
[11:49] Mike.Lorrey actually no, you have to abandon the linear motor entirely
[11:50] Mike.Lorrey because it has one parameter that should be a vector, but the compiler insists on a float, and another that should be a float, and it insists on a vector.
[11:50] Sheera Khan: the OS standard is a moving target itself... at this moment it is the LL defined protocol with it's changes now and then ...
[11:51] James.atLLOUD is wondering about the dynamics of the team that developed consensual reality.
[11:52] Mike.Lorrey James: God died, now reality is fragmented into billions of reality distortion bubbles, as this election showed.
[11:52] Andrew Hellershanks: People fork projects for a variety of reasons. They may feel they don't like something about a project and think they can do it better, or their may be bugs affecting them more directly that aren't getting fixed as quickly as they would like so they fix them, or they may be wanting to include features that the original project might not be planning to include.
[11:52] James atLLOUD: :)
[11:52] Sheera.Khan but there is only one physics in the real world...
[11:52] Mike.Lorrey unless you are talking to a subjectivist
[11:53] Sheera.Khan flüstert: until Trump declares an alternate physics someday
[11:53] Andrew Hellershanks: Um... let's not go there. :)
[11:53] Mike.Lorrey the marxists have been denying economic reality for over a century
[11:53] Mike.Lorrey lol
[11:53] James atLLOUD: @sheera - would not surprise
[11:53] Marcus Llewellyn: Running a grid is hard. One of those reason is that to do it well, you need your own dev team, be it large or small. All the largest grids have such teams. Those teams sometimes contribute back to OpenSim, sometimes they don't. This has been the reality for a very long time.
[11:55] Andrew Hellershanks: We are near the top of the hour once again. Any other OS related topics for today?
[11:56] Mike.Lorrey yes
[11:56] George Equus: A very enlightening discussion.
[11:56] Mike.Lorrey as many know, I've created the Opensim Embassy in SL to promote OS to the SL community, teach people how to migrate, etc. I'm looking for more grids to participate
[11:57] George Equus: Hour went past real fast  :)
[11:57] Andrew Hellershanks: George, it did indeed. Just like last week.
[11:57] James atLLOUD: some deep thoughts
[11:58] George Equus: I am a lay man on all this but still find subject very interesting
[11:58] Andrew Hellershanks: Mike, What is the location of the Embassy in SL?
[11:58] George Equus: Heart of the matter---- collaboration
[11:59] Mike.Lorrey its in Larsen region, one of the few builds there,
[11:59] Andrew Hellershanks: ok, ty.
[11:59] Mike.Lorrey We're having another event there on the 2nd so please tell your grid members to invite their SL friends
[12:00] Andrew Hellershanks: I'll pass it along to a few grid owners with whom I talk to on occasion.
[12:00] Mike.Lorrey thanks Andrew
[12:00] Mike.Lorrey we have weekly grid owner meetings in SLexit region of Kitely to discuss cooperating on promotion and marketing OS
[12:00] James atLLOUD: Mike, do you have a link for an event desc?
[12:00] Mike.Lorrey those are saturdays at noon
[12:01] Sheera.Khan noon of which time zone?
[12:01] Andrew Hellershanks: Ah, I was just about to ask you what time, Mike. :)
[12:01] Mike.Lorrey the event on the 2nd is at 1pm SLT I believe, I will get you a schedule later today.
[12:01] James atLLOUD: OK ty.
[12:03] Andrew Hellershanks: If there is nothing more for this weeks meeting I'll call this one done.

Personal tools
About This Wiki