Saturday, September 3, 2016

***2016 HG: Comments on OSHG Development (collection 2). HGProjects

Comments on OSHG Development (collection 2). HGProjects
Selected comments about OSHG development elicited by an initiative of Alec Genesis, along with a summary  


The following selected comments are derived from the initiative by Alec Genesis and are intended to summarize the general objectives in the discussion.  I think the main thrust is that we need and organized approach.  A previous effort (OpenSim United) to develop an organization to support that did not succeed.   

In my opinion, OpenSimulator and Hypergrid (OSHG) need such an organization now.  The organization needs to represent the main stakeholders in OSHG.  From the organization we need a roadmap for the development of the software that supports OSHG (including viewer and installations).

I believe we can raise adequate funds if we have a credible plan (roadmap) and credible support. As noted in an article below. New Media Arts want to help support this development and can provide a non-profit, audited channel.     

Copying from a G+ comment string is tricky, so I may have left things out.  Please feel free to point out important omissions in the comments section of this article.  Or you can add new remarks there.  If you want to write your own article for this blog, please contact me. 

Related articles

***********************************
  • (More after the break)
  • Where is Arcadia?
********************************************************************** 

Metaverse events, current and upcoming


***********************




Terry Ford (Butch Arnold)

As some may remember, I was part of an effort last year to put together an organization called OpenSim United. The ultimate goal was to "Unite" the OpenSim Community into an organization which was governed by the community for the community. When I say "Community", I am talking about users, grid owners, and developers (viewer developers included), commercial or not.

We had hoped the organization would be able to organize the many wants/needs of the community from all sides.. Users, Grid owners, and Developers.

We initially had many users, and grid owners interested, but only one developer showed any interest in what we were attempting.. "Justin".

Justin had said he would participate as his time allowed, but we never even got the idea off the ground.

I still have the domain name;opensimunited.com which I would gladly donate to this cause.

I also would be very much interested in being part of an organization such as this, but my time will not allow me to "Manage" it.

We had talked about collecting the wants/needs (fixes, new features, etc.) of the community, organizing these wants/needs into lists which the community could then "Pledge" support in terms of money.

These Pledges could be assigned to each task and as the community made pledges, the "Bounty" for fixing,adding, etc. would continue to grow.

Any developer could, at any time choose to "Take" the then accumulated bounty for a project by submitting approved code to that end.
Once the code was approved and tested, we as an organization wouold release the bounty to this developer or "team".

I've taken some heat over the years as I support the commercial side of OpenSim as well as the "Free" side.

While "Free" is a good thing, we have to come to the realization that at some point, we the community have to step up and "Pay" for the advances and "Fixes" if we want them in an orderly and timely fashion, otherwise we are all at the mercy of those developers who volunteer time to participate and may or may not be interested in the same things the rest of us are interested in.

I as both a grid owner and a user would be happy to participate in a system like this.

I know from my efforts last year towards this organization that there are many users and many grid owners who would be interested in the same type of organization, and while there may be some differences of opinions, I'm sure if we "ALL" were to get together on something like this, we could make it happen.

A system like this could help to motivate existing developers and/or attract new developers, it could help to define a "Path" in which we, the community would follow, it would help to advance the OpenSim project in a more organized way.

Again, this would not be a private entity.. it would need to be a totally transparent organization, non-profit status, managed by members of the community and should probably have a "Board" made up of developers, grid owners and users which would "Steer" the organization. This board in my opinion should be designated by "Vote" and each members terms should be limited, but at the end of the term they could be "Voted" back into the position if the community voted to do so.

While I think we are indeed at a crossroads, and we do risk the potential of many forks, I for one would not like to see that happen.

*****************************

Seth Nygard

From my perspective there are two possible directions such an initiative can take, each with pros/cons and somewhat different requirements. The preferred way would be to work with the existing core dev group as an advisory committee. The other way should the first prove to be not possible would be to start a new fork and manage development there, being open to input from the core devs and allowing them to join should they want.

While a fork is the not preferred and requires enough people to be able to actually do development it may be required. Historically I have seen the core dev group actively discourage people from helping with development in one way or another. One of the best public examples of this can be found if you read over the dev mailing list and look at how the MOSES group was dealt with. I for one found that whole situation to be very disappointing and I am sure I am not alone.

I have had a number of people approach me and express interest in doing development if there was a more managed, and perhaps open, way that could be handled. While this path would be a monumental task it is not impossible.

A fork however IMHO is the least preferred method as it further dilutes the work being done. Sure it may address the issue of being able to manage a road map and work from a community perspective, but it would also be starting from a major disadvantage due to a disconnect from the collective domain knowledge that exists in the core dev group.

Even to contribute to the existing core is a monumental task that is significantly complicated by the lack of, and often contradictory, documentation. The complete lack of any road map or managed structure has resulted in a fractured and inconsistent code base that any sane developer would run away from, as many have.

It is my opinion that OpenSim as a project currently sits at a crossroad, one where if the path is chosen correctly could see great things. I am however not so confident that the path I have seen with the current 0.9 dev branch over the past year is the correct path and I also know I am not alone in this opinion.

No one can forcibly change the way the core dev group operates within or deals with others. Only they can do that themselves. The ball is in their park so to speak to cooperatively work with the greater community and be part of the solution. I think the community has spoken several times to be willing to provide input but they have not been heard. And no, the existing dev meetings do little to nothing to alleviate this.

Don't get me wrong with my comments. I do appreciate and give the core dev group a great deal of credit for accomplishing what they have. There has been significant effort put into this project and we would not have what we do now without the devs and their efforts. I do however think that a more manged, or at least a less ad-hoc, dev method is needed, combined with a focus on the greater community and its needs.

I apologize in advance if my comments are taken the wrong way. They are not directed at all the devs but rather the general non-democratic and often hypocritical approach to dealing with others that I have seen numerous times. There is an apparent divide between the greater community and the existing core devs. IMHO that has to change one way or another or it is my opinion OpenSim can never evolve to be a platform it could be, one we can all use and enjoy.
************************
Talla Adam

The question was "fix things" but to be honest I've not seen any suggestion for something that is wanted or needs to be fixed since we got behind the broken LookAt() function that +Dahlia Trimble fixed for us on a community funded bounty.

Road Maps and documentation are really something the core team should be coming up with in consultation with the community but they seem singularly disinterested in really doing that or engaging the end-user community in a meaningful way at all, or at least since Justin Clark-Casey took the helm. That's probably why there are so many branches from Aurora all the way to MOSES branch and the Inworldz Halcyon project.

But we don't need more branches. We need a core team that talks to users and asks them what they want as well as vetting the contributions that are made so we end up with stable code that follows the course set by a Road Map we are all agreed on.

Moreover, this community has shown it can get behind a worthwhile objective financially if it's something that's really wanted. My view is we need a stable Opensim that takes the best of what is on offer where ever it comes from and stop the cliquishness that seems to run the show and sees off good coders.

I've probably said too much but, to be honest, I've been losing my belief that Opensim core can go much further. Perhaps one of the branches might make headway and keep the dream of a fully independent Metaverse of virtual worlds alive. One has to remember that all the other so-called advanced platforms from Sansar to High Fidelity have corporate interests and money behind them. They will be much more advanced but the worlds they generate will never be as independent as Opensim.

Opensim is worth saving for that very reason but there has to be unity of purpose to take it further. I'm sure we could have a web viewer one day and maybe even voxels as part of the structure. But we just don't have the leadership any more and certainly no direction.

***************************
Virtual Outworlding with Thinkerer

+Talla Adam
Talla: It would be nice if the core team produced a roadmap, but they have not, so others in the community are taking on that task. In any case, the resulting roadmap needs to represent the interests and advice of all (or most of) the stakeholders in the community. I will soon propose a plan for doing that. I suspect Will Burns is working toward the same objective. Let's see what happens.


Talla Adam
I agree Selby but if anyone else produces a road map how can we get the core developers to follow it. I think if you produce a road map then that is as good as branching Opensim yet again and you will need to pay developers to work on it or be willing to accept code contributions that fit the map.

Selby Evans
+Talla Adam I won't produce a road map. The OSHG community will. That will include the Dev team if they want to participate. Halcyon already has a branch and a roadmap and funding. The 2 branches are now in competition. Other grids can go to Halcyon, if they choose. If OSHG does not have good prospects of development, the commercial grids will have to consider that.
************************************

Mike Dickson

A roadmap needs to come from a functional project team with a process that allows others to contribute effectively to it. Ideally items on the roadmap are included in part due to input from the community at large. But a dysfunctional project (team) that is unable to describe at even a high level where they are going isn't going to produce the quality and improvements people are hoping for (except perhaps thanks to dumb luck, and depending on that is obviously a problem).

I am hopeful that one of the forks can step up and provide this sort of guidance and create the platform I think all of us are hoping for. 

******************************
Fred Beckhusen (aka Ferd Frederix)

I understand the frustrations. we did successfully raise the money for and fixed the lllookat bug. Lately I am finding it easier to pay for what I want, and get it compiled in my own branch. Unfortunately, I doing what we have now ... a great way to fragment Opensimulator into multiple Incompatible visions. It needs an updater to stop that fragmentation, a fix for the hairpin/loopback that haunts so many gridders. Others will have their own ideas. But the odds of any such code patches making it into opensim are close to zero. We don't even have an official branch for PhysX.

We need to find a way to hire and get more programmers for core. I had hoped Overte was the one to do it, but it is gone now, and there is no way to nudge Opensimulator in any direction other than individual self interest.

***********************************
Leighton Marjoram  Shared publicly - 12:28 PM
Couldn't put it better myself this is why I can only use virtual worlds for education currently (it is possible to demonstrate how virtual worlds could be used for therapy just not hold therapeutic sessions with clients.)

Professional practice as a counsellor would require opensim to be hardened for data security (static and in transit), confidentiality and to a degree privacy.

[Quote]
"- We have policies that we must adhere to when deploying software to our networks. You just gave a good synopsis of one of the policies governing how to deploy a database. We can't allow a database to be accessed by anyone or anything without verifying identity.

- So... if we are going to use Open Simulator or some future derivative, then it must be hardened to adhere to our security policies.


- I'm not just a drone following the rules blindly, these policies are derivatives of guidance from some very smart network security advisors from both the Dept. of Defense (DoD) and National Institute of Standards and Technology (NIST)." [end quote]

****************************

Minetheree Athanasios
I'm getting a bit lost in this conversation.

It should maybe not be overlooked that there are several forks from core around. I really have no idea if they were documented or, if done, how thorough, nor how much effort would be required.

I know of Arriba, which is being used currently by some. There is or was Aurora and also Whitecore.

If these are no longer being developed, but if they had run along with core more closely as well as maintain HG connectivity, I wonder if they can be used as a basis for further development.

They halcyon fork wants to forego the current HG connectivity until, or when, MOSES or others can create a more "hardened" one.

I don't really see those of us using core wanting to wait for some possibly idealistic hypergrid connectivity, in the meantime.

But it would seem more fruitful, possibly easier to use those other forks as the basis to build on.
*********************************************

Talla Adam

+Minetheree Athanasios 
Arriba and Whitecore, as far as I recall, are all forks of Aurora which itself marked quite a radical departure from core Opensim and gave us a faster system, var regions and npc's. Along the way Hypergrid got broken though. Revolution Smythe, the brains behind Aurora did code a bridge solution but that broke too. The dev's who work on the Aurora code branch, Arriba and Whitecore, don't seem to have any great enthusiasm fixing HG either. Inworldz took a very different route in which HG had no real place since the closed grid mentality was at the heart of it. Halcyon, built by Inworldz dev's, and MOSES work could change all that though and the Halcyon branch might yet become the preferred platform in the future.

I have to agree with what +Seth Nygard has said and he points out clearly what we have been up against, the bad treatment of the MOSES team being a case in point. I agree also we must never loose sight of just how much the core developers have done for us and our community would never have come into being without their work on Opensim. But the dev's have to consider what the community wants and not just keep heading in a direction none of us have a say in. Yes, we can get together and fork Opensim yet again so we can have a road map and much more community involvement - even financial. Businesses like Inworldz and Kitely will always keep back the code that gives their service an advantage over others even while contributing some code to core so, I think it safe to say the only way to get the kind of platform we all want in order to enjoy the best possible Hypergrid Metaverse may have to be a fork with very clear objectives writen in and over and above the business of the platform code itself. Not that the code is secondary but that the road map and objectives are never lost sight of.
*******************************************

Minetheree Athanasios

Whatever direction this all goes toward, without hypergrid connectivity NOW I have no use for it personally.

But I also will continuing doing my own little part to champion the hypergrid.

If something comes along that is hypergrid connected, but also allows for backward connectivity to those who don't keep up or those diehards who won't embrace it, then the more the merrier.

Those types have all sorts of reasons to do what they do NOW and represent the thousands of places uncounted and without representation.

In the free hyperverse many of us like it just fine and have no issues as perms are not much of a need for us, most of us are quite happy to be as altruistic as possible and share what we do.

The list would be too long to point some of them out and I would likely miss some.

But there is one single fact about the current hyperverse, and it is referring to the few iterations of commercial but open grids...that they also get to use and enjoy and enhance their grids with opensim free content.

So many of those I was in, closed, who had much, much less to offer due to being closed. Most people in those places, exclusively, have no clue at all what they are missing and those types of systems inherently choose to keep their eyes closed.

Most of all I want to see people have eyes open choices seen clearly, and that is what I personally try to do all the time.
*******************************

Fred Beckhusen (aka Ferd Frederix)
I am going to limit my response to the question asked: what would you want fixed first? My tl;dt answer would be "attract new users', 'attract new devs', and "grow the hypergrid", as follows:

Opensimulator needs to be an Application.
We need a much simpler installation. An easy-to-use program that installs, sets up, and runs Opensim, with uninstall capabilities. It could be based on a modern install tool that people are familar with, such as InstallShield, Simple Setup or Setup Factory, along with the few custom bits that set up the network and INI files. Firewalls can be automatically opened up with these tools.

It's been done before
Years ago I my first C# program was simple setup program for a charity to give out as a Thank You for donations. You could plug in a thumb drive or DVD, run one program, and log in, and ride my horse around in a Linda Kellie Western sim. But some of the tech used in it, such as the standalone version of Singularity, and the Mowes package are obsolete now. Divas Onlook viewer needs to be integrated with it instead. The user-unfriendly log-in entries for the local server is still easily solvable at installation time by making a custom shortcut con with --loginuri parameters, which this package did. I still have the source, any interested devs should contact me privately.

A strong case can be made is that a general solution needs to be in Dot Net (just like Opensimulator is) so it runs on everything, especially on Windows 7 and 10, where 85% of the world's desktops are.

Raspberry Pi Users
I have been working on a small scale solution for the six million Raspberry Pi users out there using pure Perl. It automatically sets up and connects Opensimulator 0.8.2 to OsGrid without prompts, entirely automatically. A pure Perl web server is in it so you can customize things such as the sim name, but even that is optional. Mine is set up for Linux and Pi's, which means only the dedicated hobbyists will be able to use it.

We need more people to use and support the OnLook viewer. Opensim core already has the ability to hide a lot of viewer complexity from new users. XML settings can add or remove any UI component. This needs to be documented on the Opensim wiki, and not on just Divas blog.

See http://www.metaverseink.com/blog/opensim/onlook-server-side-configuration/#more-639

It would be lovely to see these changes in Firestorm, too. But that's another set of devs to convince, and they don't care either, as no one I know here uses it, anyway. I have not had time to try it, either. But we could get together, and put together a strong group of people that do support it, and make our own social network of support for it. For example, why do we not all get together, play with it, improve it, and steer this towards much wider use in osGrid and Metro?

For one simple example, we could come up with a "User Friendly" logo and use it everywhere on social media, and at the same time make it much simpler for people to log in and have fun immediately, without all the confusion.

Loopback needs to to be fixed
Most routers aggressively refuse the hairpinning required to reach back inside the network from the hypergrid and also connect to your home server. It shuts down many people's ability to run a home server oor to expand, like me Without hairpinning, you can only run a viewer on the server. The requirement for hairpinning in your router, also known as loopback, was a bad design decision. Removing loopback is one of the few technical fixes that can actually grow the grid and possibly attract new, technical users who are familiar with Linux and compilers and C#, ( potential devs) which we badly need.

There is already code out there that fixes it for 0.7.6, but it needs to be ported upward and wider so it works in all grid configurations, and then tested and supported by core.

Remember there are six million Raspberry Pi's out there, which can easily (and cheaply) run Opensimulator for under $50 total cost for hardware plus the micro SD. But most people will never be able to use it. All because someone made a very bad design decision, and thought routers exposed private IP addresses. They don't.
*****************************************

Han Held
0.8.2 is my bird in the hand, it might eventually stop working due to bitrot, but not for a while (I expect it to be good until 2022, if not longer).

You might be surprised how active whitecoresim has been lately, though I'm not sure how suitable it would be for anything commercial. Also, no hypergrid.

I haven't tried it since 2014, but I'm reasonably sure Arribasim has hypergrid, but is it still developed?

I used bridged virtualbox instances to work around loopback issues; which worked very well for me. Might be overkill and there's no virtualbox on ARM ...but there is qemu, and it could be bridged?
********************
Fred Beckhusen (aka Ferd Frederix)
I want to add more servers ( I currently tun 3). Loopback 'sort of' works on FIOS. But it makes transport between sims iffy. Even between sims on the same computer.

Because I have to run a software loopback adapter, I also must log onto the server I want to connect to and run the viewer locally.. And that's just not possible on a Pi, or a WandBoard, or any server that does not support a GPU card. So I can only reach the machine next to me by remote accessing into my work machine and using the viewer there, which is horribly laggy. Sucks to be unable to use a WAN/LAN network as it is designed to be used. This really does need to be fixed. It's a horrible design that is unfortunatley well-baked in. Opensim knows perfectly well whether a viewer is inside out outside the netmask. It just does not care.
**********************************************


Magnuz Binder
Arriba forked off from core OpenSimulator version 0.8 Dev back in 2014-05-08, so it has hypergrid connectivity and varregions. The developer, FreakyTech, has also incorporated a few postfixes from core, but has straightened out many things regarding assets, inventory and teleporting in ways making later core fixes in these areas redundant. He is at present maintaining this fork to keep it compatible with core ROBUST and hypergrid (latest commit 2016-08-21), but otherwise working on a new, not yet published fork, "Arriba 2", which is a more thorough refactoring of the codebase, instead of just patching broken parts up.

Aurora forked off from core OpenSimulator version 0.6.9 Dev back in 2010-03-20, and WhiteCore forked off from Aurora back in 2013-12-23.

My stats regarding which OpenSimulator versions various grids are running are based on the simulator version in the landing regions of the approximately 200 grids I've been able to visit by hypergrid during August 2016, so it says nothing about the 150 "known" grids I have not been able to visit by hypergrid.
**********************************

Seth Nygard
Most things related to HG need to be fixed in one way or another. The existing HG solution is little more than a proof of concept and far from a solid working solution. Sadly when the most heavily active devs don't even use HG it gets the least attention so it is little wonder that there are ongoing issues. HG is in my opinion perhaps the single most important feature of OpenSim. "Fixing" its issues may not however be so trivial as doing so may not be compatible with what we currently have.
***********************

Dahlia Trimble
This has been an interesting read so far. I agree with a lot of the points that have been raised. Though I'm no longer involved in OpenSImulator (and have no influence any more), I thought I'd throw in a few comments to help people understand why there has never been a roadmap.

OpenSimulator started life as a open source emulation of a SL simulator, and later a grid. While there have been efforts by several devs (including myself) to make it a more generic virtual world server, these efforts have been overshadowed by the expectations of the vast majority of users that it should emulate SL as closely as possible. Later on people started setting up grids and they wanted their grids to be exactly like SL and would complain quite a bit if it deviated from whatever SL does (I'll leave it as an exercise for the reader to determine why). There is also the dependency on the various SL viewers. I've been involved in several viewer projects which attempted to move away from the SL model but it's been for the most part a very lonely effort.

The general mode of working with non-core contributors is that since we don't compensate them, we don't tell them what to do. Often potential contributors will approach a core dev and ask what needs doing and the response is usually something like "Take a look at Mantis and see if any bugs on there are something you may be interested in fixing, and join the IRC channel so you can chat in real-time with the active core developers to get any assistance you may need". Others may approach with a clear idea of what they want to do, say a new scripting engine, then, not realizing amount of work it would take to implement it, soon give up. There are also cases where personality or political issues come into play and unfortunately this does scare away some contributors. The core devs also have coding and functionality standards that they try to maintain and enforce and some contributors may not hold these standards of value and disregard them in favor of their own and this can result in conflict.

Generally contributions are accepted based on their ability to improve the software as a whole. If a contribution breaks another part, or if it breaks SL compatibility it is usually rejected. If a contribution adds some desirable feature but does it in a way which disrupts the existing code such that it makes working with it very difficult it may also be rejected or we may ask that it be modified. If a contribution is found to create a security risk, it is rejected. In some ways these can be viewed as "holding back" the project but then again there are a lot of users who are usually quiet until their favorite feature breaks. There are also a lot of businesses and institutions who depend on the existing functionality and the core group doesn't want to disrupt them. There are also developers in the core group who have fairly large scale for-profit grids and they have a lot of influence over keeping SL compatibility. It also should be noted that accepting a contribution also means providing support for it in the future, and the core group may not have the resources to support it and if the contributor also refuses to provide support, it will likely be rejected.

SL compatibility is a constantly moving target. The general implicit roadmap is to follow LL's lead and accept and encourage contributions which do so. It's also acceptable to contribute in ways that deviates from SL as long as it doesn't break SL compatibility. One should also consider that the vast majority of code in OpenSimulator is there to support this compatibility and this can make OpenSimulator a poor choice for any other virtual world architecture. It should also be noted that SL was designed some 13 years ago and the architecture which worked well back then may not be the best choice for future virtual world applications.

Anyway I don't want to disrupt the flow of this conversation. There are some good ideas being expressed and they may have a good chance of coming to fruition given sincere efforts and good diplomacy.
********************************
Talla Adam

Good to have you commenting again +Dahlia Trimble. Your expertise and words of advice are welcomed. From what you said I got the idea that Opensim could go much further if the dev's and the users were willing to move away from SL protocols to some extent and opt for a viewer that really does aim to support real advances in Opensim beyond the Second Life paradigm. This has been said before but, as you pointed out, the dev's do tend to play is safe, so to speak, for want of making the most people happy - or, perhaps making certain people with a lot of vested interest, happy. I'm not so sure the whole community are desperate to remain SL compatible. It does, on the face of it, seem so limiting.

I don't know much about the licensing of viewer code but it seems to me that, if anything needs forking, it should start with the viewer aiming for a new set of protocols that allow advances in Opensim code at the expense of breaking compatibility with SL probably although I'm sure a viewer could be run in either SL mode or Opensim mode. My view is to forget SL and go all out for an Opensim viewer, pure and simple - even a web viewer ideally.

I have the Onlook viewer which Diva Chanto has been developing, which I never really understood, and then there is the Replex viewer that the late Latif Kalifa started in order to explore Oculus Rift code from CtrlAltStudio which has also been added to a version of Firestorm. In any event I'm not sure whether Singularity V1 code is best to run with or go with a V3. That's something to talk about I guess because they all do different things like RLV which the BDSMers want and numerous other things and preferences. Personally, I want to move beyond the limits set by SL because it's probable that the Lab may slow up on new stuff for SL once Sansar is fully operational - especially if it takes a lot of people from SL.

If asked I think this Opensim community wants to advance and I am willing to put a lot of dollars on that if others are too.
**************************
Kazuko Yoshida

I might be too late with my comments - but in my very limited experience in Opensim the biggest hurdle I have had is Documentation - the opensim.orgwiki is very out of date and often leads one in the wrong direction - for new People who are wishing to start an opensim whether it is a grid or standalone then good documentation is needed

Secondly - Hypergrid is a wonderful thing - but as Carlos hs mentioned - better commuication - being able to see friend on other grids online, and I wonder sometimes if there is a better way to handle the suitcase

Third - Land and Estates is tied into a Moneytary system for some reason - I am not sure why this is - the Land change of ownership should be separate and work without having to install any Moneytary system, on any grid or standalone set up

But with all this I must say I am amazed at what can be done with Opensim - its a huge learning curve and alot of Fun!
**************************

Alec Genesis
Thank you everyone for your comments. I was highly enlightened by your responses. It seems there is a pretty big disconnect between the community's needs/wants and the Developers' Goals. I will take all of this into consideration.
*********************

Freaky Tech
Actually, the OpenSim code base has gained a lot of cruft over time. This is one thing practically most or eventually all of the forks share with varying amount.

A few of the worst things I encountered:
* Code duplication
* Varying code quality throughout modules
* Reinventing the same task several times in different modules
* Chaotic structure aka spaghetti design
* Overly complicated linkage through a mechanism called event.

Events resembles more like a multicast listener and is not really a good concept for module linkage all over besides being slower than a simple delegate (reference to a function in another module)

Magnuz mentioned that I am onto doing something new but he got one thing wrong: it is not a fork. Reason: The amount of cruft I encountered since beginning of the fork is far more than it should be for any kind of project. So basically a decision based on how much time it takes to fix vs. rewrite was the basis.

However, as it is a non-paid hobby project there is no deadline for that new thing. Instead, I have more or less some kind of roadmap what is needed to get it ready to be of good use.

The fork will continue on maintenance effort as long as I have no viable replacement.

At the moment due to what I determined throughout analysis of the code base, there is no easy handling of people that could apply as of now as helping hands. Having seen such a project, I thought "An open source project with a necessity for developer application examination? What the heck is going on here?". There is simply not enough time to do such kind of check up at the moment.
***************************

Ada Radius
+Alec Genesis The entity that's been, at least tangentially, involved with dev for OpenSim, is OSGridInc, a 501(c)(3) nonprofit corporation since 2013. It has a small amount of funding that it uses to run the osgrid.org grid, and their mission is to provide the testing platform for the open source open-simulator virtual world software. It doesn't have the infrastructure to do the kinds of fundraising that it would need to work on OpenSim dev specifically (such as filing at ~40 US state agencies that make it legal to solicit donations online, crowdfunding, some grant aps).

+Selby Evans and I are part of the New Media Arts Inc team, also a 501(c)(3) nonprofit, (and Selby does much else besides) and we do have the infrastructure to help out if that turns out to be useful. Two of our projects depend on OS (or a reasonable equivalent) working, with HG access to as many other grids as possible, so we're deeply interested, to say the least. Whether the right direction is Halcyon or OpenSim as it is now or how the core dev team develops - it's too soon to tell? Finding out what the commercial, opensource, nonprofit communities want and need, figuring out who and where the core devs are in the process - it's a place to start.

One organization that is doing opensource development and crowdsource fundraising successfully is the Blender Foundation - familiar to those of us who do 3D modeling, texturing and animation. Virtual world art is a small subset of the uses of the Blender software. The foundation pays a very few people to run the development process, run the Wiki and websites, and provide the free software. Many programmers also contribute code for free, or plugins either for free or for small fees, with much feedback from users at the forums. It works.
************************************

Fred Beckhusen (aka Ferd Frederix)
+Seth Nygard The moment you mentioned DNAT and SNAT, we lose the consumer. Until this is solved, Opensim hypergrid for the average consumer is a cloud product.

Opensim runs quote nicely with small var sims on a Pi. Good enough to fly a dragon, build a few thousand prims, and run an NPC' greeter. My thought is that if a Linux-aware person ( by definition, any Pi user) tries it for even a minute, it's a win. And a Pi is trivially easy to do given the right download. You just burn it, and boot it. But you probably cannot connect to it and also access the hypergrid because Opensim is made that way BY DESIGN. That is unacceptable.

The same build should also run on any other Linux or Windows box, ( am not a mac person, but its pure perl, so I suppose it could run there too). But even with a "real" Linux box, or Windows or a Mac, they probably cannot connect to their nifty server with a laptop without serious router and network settings.

I don't know how it can be done by an internal router. It does not seem possible without the router on the WAN side also supporting it. I have read in many support forums that I would lose my TV directory guide and all FIOS support unless I rewire the house.

The only solution I have found (for me/FIOS ) is to rewire the house for Ethernet instead of cable (done), switch the box to Ethernet from coax, which will take convincing them via a support call, and then add the other router. This configuration is unsupported by FIOS, so I will be my own tech supports person from then on. And I have to take it all down and redo it all back if I need support to watch TV again.

This is totally unacceptable in an user-installable product. The moment anything requires Linux, or a special reflashed OpenWRT router, and cannot be one-click installed and set up with reasonable defaults, it is a non-consumer product.
**************************************

Alec Genesis
+Talla Adam as per your first comment. Any views on this matter are welcome. I'm formulating things, which I can't exactly disclose at this time. But I need thorough and honest opinion in order to get a clear picture of the current state of things. If I'm going to do anything, then I'd like to see where things are. 
***************************


Seth Nygard
+Fred Beckhusen getting the described setup working is most certainly possible as I have done it for a few people to help work around limitations of the hardware provided by their ISPs. It does NOT require a router on the WAN side to make this work. The solution is 100% LAN side, but does require a solid understanding of networking and equipment setup for the various pieces of equipment. Those installs do also have working TV with guide and all the expected features from their ISP.

I am not so sure ever running things like OpenSim will be truly one-click or plug-n-play installs and setups. There is no one-size-fits all configuration and by its very nature is a server. The average consumer does not know how to setup servers, this is unlikely to change anytime soon.
***********************************

Talla Adam
Good to have you commenting again +Dahlia Trimble. Your expertise and words of advice are welcomed. From what you said I got the idea that Opensim could go much further if the dev's and the users were willing to move away from SL protocols to some extent and opt for a viewer that really does aim to support real advances in Opensim beyond the Second Life paradigm. This has been said before but, as you pointed out, the dev's do tend to play is safe, so to speak, for want of making the most people happy - or, perhaps making certain people with a lot of vested interest, happy. I'm not so sure the whole community are desperate to remain SL compatible. It does, on the face of it, seem so limiting.

I don't know much about the licensing of viewer code but it seems to me that, if anything needs forking, it should start with the viewer aiming for a new set of protocols that allow advances in Opensim code at the expense of breaking compatibility with SL probably although I'm sure a viewer could be run in either SL mode or Opensim mode. My view is to forget SL and go all out for an Opensim viewer, pure and simple - even a web viewer ideally.

I have the Onlook viewer which Diva Chanto has been developing, which I never really understood, and then there is the Replex viewer that the late Latif Kalifa started in order to explore Oculus Rift code from CtrlAltStudio which has also been added to a version of Firestorm. In any event I'm not sure whether Singularity V1 code is best to run with or go with a V3. That's something to talk about I guess because they all do different things like RLV which the BDSMers want and numerous other things and preferences. Personally, I want to move beyond the limits set by SL because it's probable that the Lab may slow up on new stuff for SL once Sansar is fully operational - especially if it takes a lot of people from SL.

If asked I think this Opensim community wants to advance and I am willing to put a lot of dollars on that if others are too.
******************************)


Kazuko Yoshida

I might be too late with my comments - but in my very limited experience in Opensim the biggest hurdle I have had is Documentation - the opensim.org wiki is very out of date and often leads one in the wrong direction - for new People who are wishing to start an opensim whether it is a grid or standalone then good documentation is needed

Secondly - Hypergrid is a wonderful thing - but as Carlos hs mentioned - better commuication - being able to see friend on other grids online, and I wonder sometimes if there is a better way to handle the suitcase

Third - Land and Estates is tied into a Moneytary system for some reason - I am not sure why this is - the Land change of ownership should be separate and work without having to install any Moneytary system, on any grid or standalone set up

But with all this I must say I am amazed at what can be done with Opensim - its a huge learning curve and alot of Fun!
*****************************

Alec Genesis

Thank you everyone for your comments. I was highly enlightened by your responses. It seems there is a pretty big disconnect between the community's needs/wants and the Developers' Goals. I will take all of this into consideration.
*******************************
Freaky Tech

Actually, the OpenSim code base has gained a lot of cruft over time. This is one thing practically most or eventually all of the forks share with varying amount.

A few of the worst things I encountered:
* Code duplication
* Varying code quality throughout modules
* Reinventing the same task several times in different modules
* Chaotic structure aka spaghetti design
* Overly complicated linkage through a mechanism called event.

Events resembles more like a multicast listener and is not really a good concept for module linkage all over besides being slower than a simple delegate (reference to a function in another module)

Magnuz mentioned that I am onto doing something new but he got one thing wrong: it is not a fork. Reason: The amount of cruft I encountered since beginning of the fork is far more than it should be for any kind of project. So basically a decision based on how much time it takes to fix vs. rewrite was the basis.

However, as it is a non-paid hobby project there is no deadline for that new thing. Instead, I have more or less some kind of roadmap what is needed to get it ready to be of good use.

The fork will continue on maintenance effort as long as I have no viable replacement.

At the moment due to what I determined throughout analysis of the code base, there is no easy handling of people that could apply as of now as helping hands. Having seen such a project, I thought "An open source project with a necessity for developer application examination? What the heck is going on here?". There is simply not enough time to do such kind of check up at the moment.
***************************************

Mal Burns

Yesterday 6:03 PM

Absolutely agree with Talla - a new viewer without that legacy Linden stuff.
And with Terry - maybe some kind of "steering committee"?

+Terry Ford
Yes Terry. I remember OpenSim United and it certainly is a good approach. I think I'm still registered with it but anyway there is also+Selby Evans New Media Arts group offering to handle the project and I noted you would not be available to manage OSU so these are good options to consider. Further to that is a less formal job by job approach is prefered then I ahve put forward a proposal for that in my reply to Siana on the Pole topic pinned to the head of OV here. Please take a look at that but all options are on the table right now and nothing is agreed anyway.

*******************
Google+ (Douglas Maxwell)

+Alec Genesis +Talla Adam +Mal Burns Let's start with the original question: "Just a hypothetical here, if a public or private entity decided to take initiative with opensim development and really fix things, what would you want fixed first?"

It has been fascinating to read through the comments here and read the inputs from such a diverse set of perspectives. Personally, I tend to agree with almost all of you!

To really "fix things", you have to start at the top with the organizational leadership...

**********************************************************************************

News and Notes

    ***********************

    The Hypergrid WIP Show

    The Hypergrid WIP is a one hour "show & tell" of works in progress or recently completed. All builders from beginner to pro are invited.  Presentations are in voice and text.  For text presentations, best bring the text in a notecard and paste it into chat.  Voice presentations may be captured in video.  Stills and videos from the show may appear in this blog and elsewhere.

    Next WIP show 

    • Next WIP show: Sun. Sept 11, Noon SLT 
    • Narasnook  location (second Sunday of the month)
    • Pandora allows presenters to run high threat OSSL functions.
    • world.narasnook.com:8900
    • Put the line above in your World Map next to Find.  Click FindTP
    • At Narasnook, use World Map to search for Pandora
    • Narasnook  location (second Sunday of the month)
    • Pandora allows presenters to run high threat OSSL functions.
    • world.narasnook.com:8900
    • Put the line above in your World Map next to Find.  Click FindTP
    • At Narasnook, use World Map to search for Pandora
    • Cookie II location (fourth Sunday of the month)
    • HG address below: paste into the World Map next to Find. Click Find, TP
    • grid.kitely.com:8002:Cookie II 
    • in Kitely: paste into Nav (top) bar of Firestorm, Enter.
    • hop://grid.kitely.com:8002/Cookie II/68/369/22

      Previous Articles from the WIP show 


      HG links-- depending on your interests 

      Radio in the virtual world

      Metaverse beginner help

      Schools in virtual worlds

      • I publish 5 or 6 days a week, skipping 1 or 2 days in midweek
      • For more on topics like this, follow Selby Evans in Google+ 
      • What do we do in Virtual Worlds? 
      • Google search this blog: Search bar, upper left--or:
      • Put site:virtualoutworlding.blogspot.com at the end of the search terms 
      • Annotated screen shots made with Jing
      • All original content on this blog is Creative Commons License, attribution only. 
      • Second LifeLindenSLurl, and SL are trademarks of Linden Research Inc. 
      • This blog is not affiliated with Second Life or anything else. 
      • Ads are from Google.




      1 comment:

      1. Alec here, I'm glad to see word has gotten around. I know that people would prefer to keep OS development in the public domain. I don't feel the same. I'm not suggesting a commercialization by any means, but many have tried in the volunteer, crowd sourcing spirit, but not a lot appears to have resulted. The lack of information on setup and administration for one is a fairly clear indication of that. Additionally, because issues still stand out, I feel that not enough has been done.

        With that in mind, I'm simply going to private investors and going straight into development in a controlled environment. This is by no means with the intent to withold new developments and fixes from the community. Quite the contrary, it is my aim to apply professional software development practices to bring opensim into maturity so that it can meet and keep pace with the growing virtual/augmented reality industry. Without swift action, I believe that it could be passed up, and there is so much potential, it can't be allowed to die in the hands of debate and indecision.

        ReplyDelete