tl; dr. Adam
From: [email protected] [mailto:[email protected]] On Behalf Of Mark Malewski Sent: Wednesday, 30 September 2009 3:45 PM To: [email protected]; [email protected] Subject: [Opensim-dev] OpenSim 0.7 Release Candidate with ALL working OpenSim Modules Diva, >> It would just be nice to get everything integrated back into core (or as >> OpenSim modules). > >This would be terrible. Diva, please explain WHY would having a working OpenSim distro be terrible? Having something that actually "works" is terrible? In my opinion, just the opposite is true. You can spend your whole life developing things (that no one actually uses, and that don't actually work or do much of anything, and that no one will ever use) or you can make a WORKING product that is usable, and that is EASY to use, and that people will use. You seem to prefer the latter. > I think you, and maybe others here, may need to understand better this >concept of extensible systems. That's at the very core of OpenSim from >the beginning, even before I started contributing -- OpenSim is not an >application, it's a platform with which to build applications. I think you need to sit down and understand the concept of "working product". You also need to stop confusing "extensible system" with "not working" product. PHP, and Apache are what I would consider "extensible systems". PHP is easily downloaded, and it works (out of the box). Yet it comes with many different modules (as part of the default distro) and those modules have all been thoroughly tested, and can easily be enabled by simply uncommenting out the module name in the default.ini file. >From an engineering standpoint, extensibility means the system is designed to >include hooks and mechanisms for expanding/enhancing the system with new >capabilities WITHOUT having to make major changes to the system infrastructure. OpenSim doesn't seem to be "extensible". OpenSim seems to be "broken". There is a big difference. Maybe your definition of "extensible" means that it requires a rocket scientist just to get the trunk to even compile (or even work), and takes hours and hours of debugging code, just to get a module to even work. That isn't my idea of "extensible". I understand over the past few months, the server infrastructure (and architecture) has been changing quite a bit. It's hard to even tell if ModRex (or any other modules) even work with the current OpenSim trunk (or latest build) at this point. The average layperson doesn't want to spend hours and hours trying to compile from source, or debugging code, or searching for plugins/modules that may (or may not) exist, and even worse many of them may not be updated, or may not even work with the current OpenSim as "core" evolves. Often times many of these modules are not updated, and most have no clue how to even build from source, and for this reason it might be good to just have VERY simple "turn-key" distributions available for download. (Stable releases) Similar to how RealXtend has done in the past. I supposed I could sit down and begin working on creating a fully configured VMWare image of OpenSim with various modules installed and configured, that people could easily download, and be up and running in a few minutes (without having to hunt for various modules, or applications), or sifting through outdated wiki pages trying to figure out how to even get started or even get up and running, but to be honest most people just want something VERY easy to use, VERY easy to setup, and would love a nice GUI interface (like WixTD, etc.) that they can use to administer the server, add users, etc. Most laypeople don't want to hire a software engineer, or a programmer, just to get OpenSim to compile, or even get a module working, or just to get OpenSim running on a machine. If I want to use a plugin with Firefox, I've NEVER had to compile or debug code. If I want to enable a PHP module, I've NEVER had to debug any code. Most modules are included in the default distro, and modules can easily be turned on and off, by simply "enabling" them in the default ini (configuration) file. In my opinion, you may be confusing "extensible system" as an excuse as to why nothing should work properly. In my opinion, EVERY single working module that exists for OpenSim should be included in the default distro (in the modules directory), and these modules should ALL be disabled by default, but can be easily enabled by simply uncommenting out ONE single line in the default.ini configuration file. Include EVERY single working module with the default OpenSim distro, so users have a list of default working modules that are regularly updated so that they actually work (and are not broken), so that when a stable release comes out, a user can just enable or disable whatever modules they wish to use (by uncommenting out a line or two in the default .ini configuration file) and those modules are in the modules directory, and can easily be enabled by just uncommenting out a single line in the default ini configuration file. The problem is, it seems like a herd of cats are headed in all opposite directions, and people really just want something that actually works. Diva, is that honestly too much to ask? There are Applications and there are Operating Systems. What do you call OpenSim? Is it an Application or an Operating System? (or is it neither?) When I say "works", I'm talking about someone can download OpenSim, and be up and running (designing things from within the OpenSim Application platform such as creating 3D content, in-world). Not sitting down and downloading source code, or attempting to figure out how to learn C# or C++ or how to write a module, just to get simple things running. The thing that made RealXtend so popular was that it was easy to use, and they had distros that were already setup and ready to use (even with a nice "beneath the sea" demo world as part of the distro). Keep in mind that most of the people interested in OpenSim as a 3D development platform are laypeople, and are graphics designers (and Second Life users) that are NOT Computer Science majors, and are not engineers, and really don't know ANY programming languages (some may know a bit of Java, or HTML, or LSL), but most don't even know C++ or C# nor would they have any idea how to even compile or build from source. They just want to use OpenSim to design 3D content, and create their own virtual world. Do you expect a web developer to know C? or C++? or C#? Try thinking of OpenSim as a "3D Web Server" for users (similar to Apache). Yes, Apache is extensible, and many modules can be written for Apache, but most of the common modules are already tested and included with the Apache distro. Modules are tested, and included with all the latest releases, and users can easily comment (or uncomment) out a single line in the default configuration file, and have the included modules working. So I believe the key to making OpenSim widely adopted as a "usable" platform for 3D developers, would be to make OpenSim easy to use (so that ANYONE can get up and running in less than 10 minutes). I believe every single tested module should be included with the default distro's. So that users can easily enable/disable whatever modules they want, and users know that the modules included with each distro have been tested, and are working modules. At this point in time, does ANYONE actually know what works, and what doesn't work? Do we actually have a working distro, with working modules (that have been tested to work) with an actual OpenSim release? Since 0.7 release is supposed to be coming out soon (in a week?) is there any way that we can stop development, and begin testing all the OpenSim modules, and add all the OpenSim modules (that have been tested and are working) to an OpenSim 0.7 release candidate? RealXtend does a very good job of doing this (with their old distros), but now that ModRex is integrated with OpenSim core, we're back to the drawing board again. If someone wants to enable ModRex, they should just be able to uncomment out a line in the default .ini file, and all the features of ModRex should work. If someone wants to enable currency, they should just be able to just uncomment out a line in the default .ini file, and now the currency module should be enabled. Why not make things SIMPLE and EASY to use? If someone wants to write a module (and wants it included with the OpenSim distro) then it needs to be tested, and once it has been tested (and confirmed to work) then it can be included with the OpenSim distro. This way at least we know what modules work (and are tested). OpenSim has evolved so quickly, that I'm not quite sure what modules even exist (or even work) at this point, and I have a few old distro's running, but I was too scared to even upgrade because everyone said that "OpenSim is currently broken" (due to all the latest changes) and people really just really want a WORKING distro (with working modules). I'm still running OpenSim 0.62 and ReX Server 0.4 on my local machines simply because it has been months where things have been completely broken (as OpenSim trunk would not even compile) and OpenSim has been making some backend changes and I'm still not even sure that ModRex/RealXtend even works since it has migrated over to OpenSim? I think your definition of "extensibility" and "extensible systems architecture" is different from mine. I believe in having something that ACTUALLY WORKS (out of the box), and extensibility means that new capabilities could EASILY be added without having to make changes to the system infrastructure. Your definition of "extensibility" seems to mean, nothing works, everything is broken, and you need to hire a software engineer just to get a few basic modules up and running. In my opinion, "extensibility" means that all the various modules would come by default with the default OpenSim distro, and they could easily be turned on (enabled) or turned off (disabled) by simply uncommenting a line in the default.ini file. Similar to PHP distro, or Apache server, or various other platforms. Either OpenSim is an Application or it's an Operating System. Since it doesn't run on bare metal, I certainly would NOT call it an Operating System, therefore I would consider it a software Application. I would consider OpenSim a 3D development platform. In my opinion, I would consider OpenSim a Server platform (software application) and you need both the OpenSim Server (platform) and a compatible Viewer to make OpenSim work. The problem is that OpenSim has evolved so much (and so quickly) that much of the Wiki documentation is outdated, no one is quite sure what even works at this point, and what doesn't work at this point. There is no list of recently "tested" modules (that are known to work with the current build/latest distro). Most "noobs" just really want a distro that they can easily download (maybe in a VMWare format) so they can just fire up a pre-configured image, and be up and running in minutes (instead of days or weeks). I'm willing to help test, and I'm willing to help with documentation, and I'm willing to even create "distros" that are easy to use (and that are tested and working) but it seems like nobody is working together. What if we just STOPPED developing, for just ONE week, and worked together on creating an actual distro? Just a working (and well tested) distro, that is thoroughly tested, that is STABLE, and that has all the OpenSim modules working with it? Then release it as a OpenSim 0.7 release. That's all I ask. Then after OpenSim 0.7 release candidate comes out (and it well tested, and all the modules from the OpenSim GForge are tested to work and be compatible with the 0.7 release, and then we wrap everything up, and release it as a working distro! Just halt development for 1 week, and just focus on bug fixes, and getting the modules to all work so we can just have a nice OpenSim 0.7 release candidate, with lots of working modules (that are all tested) and are included in the default distro. People can still choose what modules they wish to enable, but at least include all the known working modules with the default distro (or create a "vanilla" distro, and a "full distro" with the OpenSim 0.7 Release). That way one has the working modules, and the other doesn't have the working modules. But this way at least we can have an actual TESTED release candidate, that has all the working OpenSim modules (with updated documentation). I'm willing to help with documentation, and testing, but I just want to see an actual release candidate (with working modules) come out. On Wed, Sep 30, 2009 at 2:03 PM, <[email protected]<mailto:[email protected]>> wrote: Mark Malewski wrote: > It would just be nice to get everything integrated back into core (or as > OpenSim modules). This would be terrible. We're going in the opposite direction, which is to have a minimal core and let people do their own extensions as they wish, hopefully replacing the heck out of the reference implementations. I think you, and maybe others here, may need to understand better this concept of extensible systems. That's at the very core of OpenSim from the beginning, even before I started contributing -- OpenSim is not an application, it's a platform with which to build applications. Some extenders of OpenSim may want to get together and try to make their extensions work with each other. That's great and desirable. But let's not prevent innovative ideas from emerging by throwing a massive feature-full application out there as "OpenSim". Diva / Crista _______________________________________________ Opensim-dev mailing list [email protected]<mailto:[email protected]> https://lists.berlios.de/mailman/listinfo/opensim-dev
_______________________________________________ Opensim-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/opensim-dev
