Opensim apparently is for Firestorm only

[Geir Nøklebye]

It has been made abundantly clear by Opensim developer(s) that Opensim is developed only to work properly with Firestorm, and they cannot be bothered to test with any other viewer.

Based on the above I have retracted the Linux server resources I have provided to the developers the last 3 years, and that was used for Opensim development.

In terms of Dayturn development, it remains to be seen how well this can be worked around, but Dayturn was intended to be a light weight alternative to the bloathware that is Firestorm without excessive baggage that weights the viewer down.

It looks like the only mission statement of Opensim for the time being is to make it work with Firestorm, and whatever is piled on to that viewer either by demand of SecondLife compatibility or the Firestorm team's own additions.

Much of the SecondLife functionality is either not supported by Opensim (Marketplace, Pathfinding, Experience Keys, their help system, Voice Morphs), others are not needed because OpenSim is not running in a single instance with assets in a CDN (content delivery network), others mandated by US legislation are not needed in many (most) of the geographies Opensim grids are installed in. In addition there is the need to support Hypergrid and standalone Opensim installations in addition to VAR support, not possible in SecondLife.

Having said that, all possible credit shall be given to the SecondLife viewer development team which works tirelessly to improve the viewer code. Code without which, any Opensim viewer, including the Dayturn viewer, would not be possible.

 

[Tampa]

You and I know that's not true.

If you hadn't stormed off you would have caught what was being said. You may want to read back the rest of the meeting minutes.

To re-iterate:
OpenSim is being developed to handle packets sent by the viewers in a proper manner, no matter which viewer that is. As per design OpenSim attempts to be as compatible as possible with SecondLife, but maintains a position of diverging on parts of the technology and design where it makes the most sense to do so. A good example of that is how BoM is being done on the viewer end, which usually feature much more powerful CPU and GPU than can be provided via the server side.

We have treated your project with nothing but respect and encouragement. I have listed Dayturn as equal to Firestorm, because despite everything I still think the project has merit and pushing for fixes and changes in the viewer space, competition one might say, drives innovation. I have offered help numerous times not because "it's the right thing to do", but because I want to help and offer my free time and resources to a project that is openly interested in dialog and shows to the meetings regularly to discuss topics. I really don't understand why you would even consider OpenSim would be developing for a specific viewer or even directly Firestorm when they have only recently begun more active development towards OpenSim at all. Why you think I registered here and made posts asking for changes to go towards OpenSim specification if I didn't see this project as the first in a while actively looking to build a good relationship and compatibility with OpenSim.

I asked you months ago what your mission statement was, what the Dayturn project was aiming for and I did not really get a clear answer nor have your actions really gave me an idea on what you plan for it. As I stated before, I would more than welcome an OpenSim-first Mac-first-Windows-later viewer as that is more than needed to future drive development of OpenSim as a platform. If that is not the aim for this project, that's fine, but don't for one second think that OpenSim does not welcome a new viewer run by a competent developer.

Again, we have treated you with nothing but respect, were and still are, happy to have you at the meetings to discuss topics and developments. It hurts me to hear that you feel mistreated when that is by no means anyones intention and I hope you'll see that too. We invite you to attend the meeting next week and air your frustrations as to what you feel needs to change. Of course we cannot make promises to change any of it, but it should be widely known by now that reasonable argumentation does yield logical changes in OpenSim, at least that has been my personal experience with the project. If you do not share the same sentiment and have grievances then please share them as that is the only way we can even attempt to do something about them, communicate communicate communicate.


I am writing this with a sincere hope that you understand that if you came away from this with the idea we only develop for Firestorm that this is not the case and we are more than willing to work with you and Dayturn on making the metaverse, our metaverse(bugger off Facebook), a better place for everyone.


P.S. I know he won't like this but it has to be said, do not pay much attention to misterblue or what he says. It was a statement totally uncalled for and I got a bone to pick with him anyways so just ignore him entirely.

P.P.S CDN assets are something that is possible and have done that already, just doesn't really offer any performance increase over just a decently setup and programmed asset server. Nothing I could measure anyways.

 

[Geir Nøklebye]

MB had nothing to do with it. He has always been this way since I helped him get BulletSim for macOS compiled back in 2014 and lent him server space to continue developing it alongside the other platform versions.

It has to do with a developer who never bother to even install the viewer to test with opensim development. He might do from time to time but then there are multiple versions between, meaning months between. While at the same time I repurposed a server completely for his use to develop the mono version on.

Enough of that…

The original reason for starting development of Dayturn was that the other viewers frankly runs like shit on macOS, and they still do, because they are all second thoughts, and seen as a PITA by their developers (which also is true for LL).

The only reason why there was a Windows version was that enough people asked me to also make one, and I started (grudingly) 2 1/2 years after work on the macOS version commenced to apply all the changes that had gone into it onto what was a somewhat flaky Windows codebase. This while there was a barrage of changes from SL that needed to be integrated. Unfortunately the two never caught up code-wise, while functionality was closing in.

In the midst of this I had heart surgery with a long recovery, and very adverse reaction to the Pfizer vaccine which almost ended me.

As you know at the end of last year the Windows version became unstable and I blamed it on Windows, but the truth lay elsewhere.

Rather than try and bring the code bases for 64-bit macOS and 32-bit Windows to catch up with each other while at the same time merging EEP, VS2017 and 64-bit Windows, I re-forked Kokua which had already done this and have spent the last 2 months reapplying a large number of my changes to that codebase.

This is nearly complete, and is shortly at the point where that base will be forked to bring back the opensim specific code. The result is fast and stable 64-bit versions for macOS and Windows running the exact same code (platform differences excluded) which you can download preview versions.


My reaction is also colored by the fact that there is no evidence of any work at all being done to bring the viewer code forward to run native on Apple Silicon. Most likely at the WWDC in June, Apple will pull the last Intel based machines from marketing and shortly after OpenGL, after which is over for the viewers on macOS. – This frustrates me to no end, because a rewrite of the code is a 6-10 man year project and far beyond my capabilities and knowledge.

After that – for me, it is over. I will never invest in anything that has a dependency on Microsoft. Ever.

It also frustrates me that Opensim has no roadmap at all. Just continuous firefighting to make it work with Firestorm, which being a SL focused viewer is driven by the changes imposed by Linden Lab as they develop their business.

Now, of course, Microsoft has pulled a stunt on us all and for all practical purposes cancelled mono, which most likely will relegate Opensim to run on Windows only given that future version of .Net support on other platforms will be flaky to non-existent. So you can circle back to the point on Microsoft above.

 

[Tampa]

That the future looks dire for OpenSim is nothing new, this has been a status quo for a decade almost, but work is still being done regardless, because there are dedicated people that still see merit in the technology.

It is a bit unreasonable to expect developers to test every viewer for regressions as changes are being made, hence why we are happy to discuss them with the viewer developers when they do come up. There are, by their own accord now 5 viewers in development looking to bring full compatibility to OpenSim, testing them all would be a full time job, especially as changes are being made to some fundamental things like animations and agent updates.

A project of this scale can't have individuals be aware of all facets of the codebase on the server and client side so distributing the burden of testing and development across multiple people specializing in the areas they know most about does make more sense. If there are things in the viewer code OpenSim isn't handling properly or causing issues we have many channels to bring those up and get them resolved and that has happened quite regularly and positively in the past.

Like I said, if your focus is Mac and you'd rather spend the time there, that's totally fine. Having a dedicated Mac viewer without a Windows version is fine. Maybe one day someone comes along that wants to take that part, but no one is forcing you to do anything just because they whine the loudest. I'd love to help test the Mac version even if getting hardware wasn't a decision of whether I'll eat for a month or not(yes I know age old Mac is overpriced, but value only got better very recently).

I have offered and tried to build a system of developers to look at the integration with OpenSim as part of a general support and mapping project to outline what the OpenSim specification actually is, but as you know just as well getting developers for that to work for free and spend days upon days reading into code written ten years ago by people who still haven't learned to code is not a proposition many would take unless they have a particular interest in the project.

I can sympathize with the RL struggle as I have never had a shot not knock me out cold for a week or two and my last round fogged my brain up for a good month before I could think straight again. That's still something I prefer over some of the diseases you'd get otherwise and I can assure you your absence was noted and there was plenty of discussion as to your condition and worry about you. Develop on cold hearted machines all day throwing compile errors without regard to your feelings the humans around us become the only source of empathy... uh that went dark, sorry.

The frustrations you have with this are valid and I can echo a lot of them as well, see postgres mono 6.12 debacle and the audacity that is MS pushing unsafe code handling to nuget. It's a sad state. We are overdue for new anti-trust against MS to stop taking over half the internet and letting, what appears to be total rookies or lunatics, make decisions that are detrimental to so many communities and projects. If there is one thing I have learned though is that if you want something done properly you sometimes just have to do it yourself and show the relevant parties how they are wrong, only that tends to teach humility to some "Rockstar coders".

But I am getting off track. No one in the OpenSim project is holding anything against you or acts in a deliberate manner against you. Actions taken that you may perceive as negative and that, without context or knowledge of the character of the person that said it, may seem negative, but they are not. A lot of people can be a bit difficult to work with, that's not something I can deny as I have come to realize that myself and I know how frustrating it can be to push a million pounds of idiot up a mountain. I still value the interaction I have with the project and apply my own "dick moves" to retort the sometimes less courteous responses I get with sarcastic behavior. I'm not perfect by no means and clashing with people is something I have a lot of experience in, you know that too as I have told you in the past.

What I value in most people that I interact with a lot, yourself included, is that logical arguments tend to create results. It's not always possible to find a compromise or to convince someone of a change when they have their own plans, we are all equally guilty of that. Though there is always a path forward and if you still see any value in the project of a viewer for it then I'm more than happy to support that were possible. At the same time I can understand the sentiment of wanting to move on to greener pastures that bring more joy and I can respect that decision as well. If your feelings tell you to change your path away from the project then I can only say that we will be sad to see you go, but hope that your new ventures will bring you the joy you seek, because life is too short to not grab the bull by the horns.

 

[Geir Nøklebye]

The difference is that my viewer has been released with almost monthly releases since 2016. It is not in testing as some of the other viewers you referred to.

The heart was unstable for 6 months because of the vax with multiple calls to emergency services and hospitalization. You guys generally had no idea what was going on.

 

[Geir Nøklebye]

I have provided a few answers for you since there seem to be confusion per the Opensim Dev meeting chat log of yesterday.

Andrew Hellershanks: I don't think it was the many references to firestorm. I think it was the comment that made it sound like firestorm is the official viewer for OpenSim.

That is mostly correct. There are other factors too.

Ubit Umarov: sad is that while we where here, is did block my access to his machine, where i was running a test sim.

This is a server that has been provided by me for opensim development for around 3 years.

Given the above I did not feel it was particularly important for me to provide this service any longer, so I disabled the firewall rules that gave Ubit access to the system, shut down the opensim instance and logged out the user that owned the processes.

Despite the rules being disabled I noticed Ubit was able to start 2 new sessions on the system, so I set the account to noshell and logged out the instances in order to check if there were possibilities of gaining shell I did not know about.

Then I went on to back up the home directory with (resumably) all the opensim files in addition to make an export of the database he had used. Zipped those up, and moved the zip files to a location from which Ubit could download them.

I noticed the zip of the home was unexpectedly slim, and on inspection it became evident that all the content in that directory structure had been deleted; not by me.

Ubit has been offered the export of the database, which on restore can be used to recreate the test regions exactly as they where given all his assets are stored on the OSG asset servers.

The server has since received a much needed OS upgrade, but Ubit's account and the database is still there.

Ubit Umarov: He also did provide some help, for exmple compiling ubOde for macs, and also testing ahd patchin pgslq code

All macOS dylibs in current use in Opensim has been compiled and provided by me.
These are 64-bit Intel dylibs.
I have also compiled and provided 32-bit ARM libs of the same for use on 32-bit Raspberry PI OS

Vincent.Sylvester @hg.zetaworlds.com: I had high hopes for Dayturn as a viewer with a focus more on OpenSim than SL, but it's not clear what the goal is now

There is no change in goal. I just register that viewer development beyond Firestorm is of no interest to the so-called Opensim dev team.

A version 3.0 of the viewer is currently in development which will both bring code parity to the macOS and Windows version, in addition to adding functionality like EEP to the viewer, which proved to be very hard to add to the codebase with major effort on my side.

The new version, right now, can only log in to Secondlife, but that will change when the Opensim code is reapplied to it, and it will no longer be able to log in to Secondlife.

There is no plans to release a viewer for Secondlife, both because it would be in competition with Kokua, which it is re-forked from, but also because the I don't have any particular interest in re-investing in Secondlife. I still keep a small presence there, which pays for itself.

Andrew Hellershanks: I just looked at the Alchemy viewer website and found it is for Windows only. I looked at the Dayturn website and it is for Mac and Raspberry Pi.

Incorrect. There never has been a viewer for Raspberry PI, and possibly never will.
The Dayturn viewer 3.0 will be provided as 64-bit macOS and Windows versions.
Current versions are 64-bit for macOS and 32-bit for Windows.

 

[Tampa]

There is no change in goal. I just register that viewer development beyond Firestorm is of no interest to the so-called Opensim dev team.

To repeat, that is not the case. Not sure how many times you want me to say that, but if you read the meeting minutes you read that just as well from Ubit himself so I don't get why you are still worked up over something that isn't the case.

If you want to talk about viewer development and protocols you can attend the meeting like you did in the past and things can be worked out and fixed. I don't give zero fucks over interpersonal squabbles, I want OpenSim and viewers to function properly and everything operating in unison, that's my standpoint. You can have your own lookout on things and ideas you'd like to follow, but the accusation of only developing for a specific viewer is bullshit and you know that.

I really don't get where you even got the notion of favors being pulled for any viewer, I really don't. I went back and read the minutes again and I can't find it. Feel free to be annoyed or something, but don't accuse the project of something it did not state, discounting misterblue obviously.

P.S. We are meant to be wise adults, grown up and reasonable, this just makes everyone look bad.

 

[Geir Nøklebye]

To repeat, that is not the case. Not sure how many times you want me to say that, but if you read the meeting minutes you read that just as well from Ubit himself so I don't get why you are still worked up over something that isn't the case.

This is an impression that has been built up gradually over the last 2 years with practical interaction with the "development team". It is virtually only Firestorm that is tested, Firestorm functionality that is regarded as reference, and the only "development plan" is to make Opensim work with the next version of Firestorm.

You can read minutes till you get green in your face, but it is the day to day interaction that builds the reality of what is happening.

 

[Tampa]

If you want to be a pouting child about the fact that people mainly use the viewer they are comfortable with or doesn't throw bugs at them then that's on you. Don't expect ordinary users to be good testers.

OpenSim develops for the protocol, if viewers add support for changes to that they can implemented and an aim is placed to figure out a good way to handle the protocol. That is done for any viewer and any protocol change. It's been like that for as long as I can remember.

If you feel different about that, that's on you at this point. Everyone has made their intentions clear, in writing, for everyone to see. It's the last thing I will say on this, because this is starting to just be silly.

 

[Geir Nøklebye]

OpenSim develops for the protocol

The Protocol!

Where is that defined? – In the Firestorm source? Because I have surely not seen much of it documented for any recent versions of Opensim.