hi;
I am heavily imfluenced from the metaverse meeting in SL last tuesday, where i believe we have discussed to engage in preparing a distributed croquet grid for testing purposes. i'm actually not sure whether such setup already exists, so please tell me where to jump in just in case. Would following roadmap make sense: 1.) setup a headless croquet server for an (empty?) island what specs would such a server need to fulfill ? 2.) connect the island to other islands in the grid which topology would be adequate here ? 3.) get some more fancy avatars into the world. how can we get our SL representations into Croquet? I know this is already possible somehow, but following this list there seem to be some barriers. Can someone make a really good step by step documentation available ? 4.) export SL sims to an external represeantation. actually i am working on an export script which eventually creates an XML file containing the prim descriptions of parts of a grid node (maybe even the whole grid node) I have been talking to Lotte Twilight after our SL meeting and she eventually convinced me to put some effort into this ;-) BTW if someone allready got it, just tell me. i'd appreciate to save my little free time for other stuff. 5.) Import the data created by step 4 into the local croquet island I have virtually no idea how this could be done. So maybe anyone of the experts could contact me in order to negotiate what will be needed within the XML file i hopefully can provide. And whoever feels enlighted could provide such an import script ? 6.) Set up a place for our future Croquet meetings in Croquet. 7.) See what happens... Is this a reasonable roadmap ? Can we achieve this with the current SDK ? If this is reasonable, then i realy would like to get a good description how to setup the headless server on a linux server for dummies and what to do in order to allow external connections to the grid point (ports to open etc.) Maybe it makes sense to start writing some admin/user manuals for opencroquet and place them at a prominent place, so that interested people with lower skills still can jump in... thank you for any help. regards, hussayn |
Hi,
Hussayn Dabbous wrote; >4.) export SL sims to an external represeantation. actually i am working on >an export script which eventually creates an XML file containing the prim >descriptions of parts of a grid node (maybe even the whole grid node) I >have been talking to Lotte Twilight after our SL meeting and she eventually >convinced me to put some effort into this ;-) >BTW if someone allready got it, just tell me. i'd appreciate to save my >little free time for other stuff. Regarding item 4; as far as I understand you are trying to get object descriptions out of SL via XML, and eventually import them to Croquet. The logic behind this is to use SL for creating 3D objects. In my view to achieve a better interoperability between platforms, we should use an independent (and better) 3D authoring tool rather than SL. Assets that are created with this tool can be imported in both platforms via adapters. I found out there are scripts on Blender to import 3D objects (infact there is an open source add-on on Blender to use SL prims for creating 3D objects). If anybody knows any tool doing the same for Croquet, Blender may be the common 3D authoring tool for platform independence. BTW I see that there will always be a need for transferring objects from SL to Croquet or vice versa, so the effort you put in is definetely not a waste of time but I only wanted to discuss what should be the general direction to achieve a broader Interoperability. Regards, Kaan -----Original Message----- From: news [mailto:[hidden email]] On Behalf Of Hussayn Dabbous Sent: Friday, February 23, 2007 10:24 AM To: [hidden email] Subject: [croquet-dev] what about setting up an opencroquet grid ? hi; I am heavily imfluenced from the metaverse meeting in SL last tuesday, where i believe we have discussed to engage in preparing a distributed croquet grid for testing purposes. i'm actually not sure whether such setup already exists, so please tell me where to jump in just in case. Would following roadmap make sense: 1.) setup a headless croquet server for an (empty?) island what specs would such a server need to fulfill ? 2.) connect the island to other islands in the grid which topology would be adequate here ? 3.) get some more fancy avatars into the world. how can we get our SL representations into Croquet? I know this is already possible somehow, but following this list there seem to be some barriers. Can someone make a really good step by step documentation available ? 4.) export SL sims to an external represeantation. actually i am working on an export script which eventually creates an XML file containing the prim descriptions of parts of a grid node (maybe even the whole grid node) I have been talking to Lotte Twilight after our SL meeting and she eventually convinced me to put some effort into this ;-) BTW if someone allready got it, just tell me. i'd appreciate to save my little free time for other stuff. 5.) Import the data created by step 4 into the local croquet island I have virtually no idea how this could be done. So maybe anyone of the experts could contact me in order to negotiate what will be needed within the XML file i hopefully can provide. And whoever feels enlighted could provide such an import script ? 6.) Set up a place for our future Croquet meetings in Croquet. 7.) See what happens... Is this a reasonable roadmap ? Can we achieve this with the current SDK ? If this is reasonable, then i realy would like to get a good description how to setup the headless server on a linux server for dummies and what to do in order to allow external connections to the grid point (ports to open etc.) Maybe it makes sense to start writing some admin/user manuals for opencroquet and place them at a prominent place, so that interested people with lower skills still can jump in... thank you for any help. regards, hussayn |
I hope a lot of object editor conversions are provided.
SL's 3d representation is deeply impoverished and hugely non-standard and woefully inexpressive. Reminds me of 1960's 3d graphics modeling. That said, I suppose import into Croquet is great. The other direction is pretty much impossible (walls and spheres and tubes, yeah, meshes and so forth nope). But more importantly, Croquet is a dynamic rendering environment. Any picture that can be drawn by an algorithm can be drawn in Croquet. Static data is just input to those algorithms. Don't get too wrapped around the idea of "standardizing" around a data-centric representation. The representation that is universal is a program. |
ala Spore?
;-) |
In reply to this post by Hussayn Dabbous-2
I love a good straight line... On Feb 23, 2007, at 2:24 AM, Hussayn Dabbous wrote: > hi; > > I am heavily imfluenced from the metaverse meeting in SL last > tuesday, where i believe > we have discussed to engage in preparing a distributed croquet grid > for testing purposes. > i'm actually not sure whether such setup already exists, so please > tell me where to jump in > just in case. > > Would following roadmap make sense: > > 1.) setup a headless croquet server for an (empty?) island Makes sense to me. :-) See http://croquetconsortium.org/index.php/ Croquet_Collaborative > what specs would such a server need to fulfill ? I think the pipe is the gating factor. I ran WiscWorlds (a predecessor to the Croquet Collaborative) from a Dell desktop, but it was on a high speed university pipe. Right now the Collaborative is on a single processor Xenon server that also runs the www.croquetcollaborative.org site and ALL the xRFB servers. In fact, I usually run TWO routers and two complete sets of worlds on the same box. (One set for development.) The box has no graphics or audio and runs FreeBSD Unix. But most of the demos, and even the KAT generate a lot of traffic. Add a few people running video and audio we start to run into hosting bandwidth limits. > > 2.) connect the island to other islands in the grid > which topology would be adequate here ? Makes sense to me, too. There are two approaches already existing (in public code, anyway). See http://croquetconsortium.org/index.php/ WAN/LAN%2C_Connecting_and_Discovery Roadmapwise, note http://croquetconsortium.org/index.php/ Croquet_Collaborative#Community > > 3.) get some more fancy avatars into the world. how can we > get our SL representations into Croquet? I know this is already > possible somehow, > but following this list there seem to be some barriers. Can > someone make a really > good step by step documentation available ? Yup. How about step-by-step movies? Mark did a terrific job of this at http://croquet-bento.blogspot.com/2006/06/better-avatars.html The movies cover getting material from Linden, but not directly from SL (your step 4). (I've followed along with the movies using Poser 5. I haven't tried 6, and I haven't tried in the KAT.) > > 4.) export SL sims to an external represeantation. actually i am > working > on an export script which eventually creates an XML file > containing > the prim descriptions of parts of a grid node (maybe even the > whole grid node) > I have been talking to Lotte Twilight after our SL meeting and > she eventually convinced me to put some effort into this ;-) > BTW if someone allready got it, just tell me. i'd appreciate to > save my little > free time for other stuff. > > 5.) Import the data created by step 4 into the local croquet island > I have virtually no idea how this could be done. So maybe > anyone of the experts could contact me in order to negotiate > what will > be needed within the XML file i hopefully can provide. And whoever > feels enlighted could provide such an import script ? > > 6.) Set up a place for our future Croquet meetings in Croquet. How about the Collaborative? But I've got to find some time to make the darn thing reset on error so that folks don't keep getting nothing but a red screen after someone else gets an error. > > 7.) See what happens... > > > Is this a reasonable roadmap ? > Can we achieve this with the current SDK ? Noting that SDK means Software Developer's Kit, hell yeah! It's mostly been done, and the rest is just a matter of programming. But if you're looking for a pre-built application that does what want, you won't find it in the SDK. > > If this is reasonable, then i realy would like to get a good > description how to setup > the headless server on a linux server for dummies and what to do in > order to allow > external connections to the grid point (ports to open etc.) Ah, but that's a contradiction in terms. Setting up a server is not for dummies. That has nothing to do with Croquet. Hosting Croquet is easy. I spent FAR more time getting a server built and configured, with SSH and VNC so I could administer it remotely, Apache/PHP/MySQL/ Wordpress to explain it, jailed VNC so people could run Web browsers in-world, and learning the mysteries of /etc/rc.conf so that I could log out. But the Croquet part is trivial. On Unix, we start (inside of a /usr/ local/etc/rc.d script and with full pathnames) with: squeak -headless imagename startupscript.st and startupscript.st is basically: KMessageRouter runRouters: 5910 serverName: 'someName' serverPassword: 'somePass' log: false. KStandardHarness new setupUser: 'someName' password: 'somePass' entryWorld: WisconsinWorld address: '127.0.0.1' port: 5910; syncAllWorlds. > > Maybe it makes sense to start writing some admin/user manuals for > opencroquet > and place them at a prominent place, so that interested people with > lower > skills still can jump in... I'm working towards a no-manual way for people to fire up and connect a world of their own to the Collaborative. But dealing with machine maintenance or hosting issues is still going to be up to you. > > thank you for any help. > regards, hussayn > |
Howard & everyone,
Perhaps, Howard, your efforts could best be preserved for those who want to get a server started quickly, in an Amazon Machine Image (AMI) for Amazon's EC2. For those who are serious about starting a Croquet grid, the initial costs for Amazon EC2 & S3 are very minimal. The Croquet Consortium could give away the AMI or give it for a small donation. A small company could even sell shares of the server to reduce the cost even further. Once everyone interested can get started quickly, and begin sharing quickly, code, media, content, etc., then the more quickly we might bring Croquet to the "application" level. Cheers, Darius |
On Fri, Feb 23, 2007 at 11:33:10PM -0800, Darius Clarke wrote:
> For those who are serious about starting a Croquet grid, the initial > costs for Amazon EC2 & S3 are very minimal. The Croquet Consortium While EC2/S3 are nice, they're by no means the cheapest on the block. Some of us have underutilized hardware sitting in the rack to burn. I'm looking forward to the 1.0 SDK -- the last beta was too buggy to run, starting with cr/lf in the shell script and missing libuiid (the same issue as on Ubuntu 6.06 LTS last time I tried).. Anyone running Croquet headless on a Debian Etch system? > could give away the AMI or give it for a small donation. A small > company could even sell shares of the server to reduce the cost even > further. -- Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org ______________________________________________________________ ICBM: 48.07100, 11.36820 http://www.ativel.com 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE |
Hello,
For those with underutilized hardware, packaging the whole Croquet setup as a software appliance may be a good fit: http://wiki.rpath.com/wiki/Application-to-Appliance see the MediaWiki appliance for an example: http://www.rpath.org/rbuilder/project/vehera-base/ "An appliance that bundles up PHP, MySQL, Apache, MediaWiki and the rPath Appliance Agent to provide a self-contained, turn-key, updatable Wiki appliance." To me it looks like you can package a working installation that only contains the components that are absolutely necessary, a bit like a custom Linux distribution with completed installation. Maybe it can be (or is already) combined with virtualization, so you could run a Croquet appliance alongside other stuff on the server. -- Matthias |
In reply to this post by Eugen Leitl
If someone can put together a VMware virtual appliance for it, it
would work much better I should think. -Kyle H On 2/23/07, Eugen Leitl <[hidden email]> wrote: > On Fri, Feb 23, 2007 at 11:33:10PM -0800, Darius Clarke wrote: > > > For those who are serious about starting a Croquet grid, the initial > > costs for Amazon EC2 & S3 are very minimal. The Croquet Consortium > > While EC2/S3 are nice, they're by no means the cheapest on the block. > Some of us have underutilized hardware sitting in the rack to burn. > I'm looking forward to the 1.0 SDK -- the last beta was too buggy to run, > starting with cr/lf in the shell script and missing libuiid (the > same issue as on Ubuntu 6.06 LTS last time I tried).. > > Anyone running Croquet headless on a Debian Etch system? > > > could give away the AMI or give it for a small donation. A small > > company could even sell shares of the server to reduce the cost even > > further. > > -- > Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org > ______________________________________________________________ > ICBM: 48.07100, 11.36820 http://www.ativel.com > 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE > -- -Kyle H |
On Sat, Feb 24, 2007 at 03:10:04AM -0800, Kyle Hamilton wrote:
> If someone can put together a VMware virtual appliance for it, it > would work much better I should think. They're convienient, but the overhead for a VMware guest or Xen is rather high. I prefer to run things in leaner virtual environments, such as the Linux VServer. -- Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org ______________________________________________________________ ICBM: 48.07100, 11.36820 http://www.ativel.com 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE |
In reply to this post by Howard Stearns
Howard Stearns schrieb:
> > I love a good straight line... > ... >> >> If this is reasonable, then i realy would like to get a good >> description how to setup >> the headless server on a linux server for dummies and what to do in >> order to allow >> external connections to the grid point (ports to open etc.) > > Ah, but that's a contradiction in terms. Setting up a server is not for > dummies. That has nothing to do with Croquet. Hosting Croquet is easy. I > spent FAR more time getting a server built and configured, with SSH and > VNC so I could administer it remotely, Apache/PHP/MySQL/Wordpress to > explain it, jailed VNC so people could run Web browsers in-world, and > learning the mysteries of /etc/rc.conf so that I could log out. > > But the Croquet part is trivial. On Unix, we start (inside of a > /usr/local/etc/rc.d script and with full pathnames) with: > squeak -headless imagename startupscript.st > and startupscript.st is basically: > > KMessageRouter runRouters: 5910 > serverName: 'someName' > serverPassword: 'somePass' > log: false. > KStandardHarness new > setupUser: 'someName' > password: 'somePass' > entryWorld: WisconsinWorld > address: '127.0.0.1' > port: 5910; > syncAllWorlds. > When i asked about an installation instruction for dummies, i focussed on Croquet, not on the technical aspects of server administration. I basically have no idea about what i should do in order to setup a Croquet world on a headless server. i'll have to read the pointers you have published in this thread... I hope i can find an instruction about what i need to install, which technical preconditions must be fulfilled (e.g. open ports on the firewall) etc.. see, we have good linux admins here. all i need to do is telling them which software package to install and off we go. but there is my problem. i dont know the entry point. If someone would provide the basic cookbook like: 1.) download the croquet package from ... 2.) prepare your headless server by doing this and that ... 3.) create an empty world by doing this and that 4.) start the server and take care about the ports etc. 5.) connect from an croquet client as follows 6.) connect your world to other worlds like this. 7.) go and test your installation. if you can walk through all connected worlds step one is done. if not ask the croquet-dev list I would appreciate to see such a thing as kick off manual. If we had it, i believe we can very quickly get up a significant amount of interconnected worlds. The hype is there, SL has created large interest. Now lets get up and jump in... If we can get MANY individual setups interconnected, we could very quickly get really huge. And we will see whether the concepts can support that... If we had a popular download button [download your grid point server] which contains everything (including the installation shell script) admins could fire up their grid point in less than 5 minutes, that would be the beginning of a shockwave ;-) And if we could even add a bunch of selectable island meshes like SL does, then IMHO it would rock... If we can get to this initial point, i believe success is near ;-) see, what i mean ? >> >> Maybe it makes sense to start writing some admin/user manuals for >> opencroquet >> and place them at a prominent place, so that interested people with lower >> skills still can jump in... > > I'm working towards a no-manual way for people to fire up and connect a > world of their own to the Collaborative. But dealing with machine > maintenance or hosting issues is still going to be up to you. > sure, but as i pointed out above it makes virtually no sense to force people getting experts in Croquet before they can setup a server. that would distract people. OTOH telling them that setting up the server can be done in 5 minutes, that would bring people into the game... regards, hussayn |
In reply to this post by Eugen Leitl
Not everyone has the luxury of a VServer-based virtualization
environment. Also, not everyone has the time necessary to learn how to administer one. If you want to continue with Croquet being a high-cost-for-entry development environment, please be aware that you will merely cause more and more people to be frustrated by the difficulty and veer away from the promise that Croquet has to offer. If you would like to make Croquet more easily accessible, please be aware that consumer- level/non-OS-dependent virtualization is a low-cost/low-barrier means to do so. If I can download a virtual machine that runs on Linux, configures itself, and runs a headless server so that I can see how the client interacts with it without having to have a separate machine, I'm more likely to be wowed by it than by something I have to wrestle with just to get only a red screen out of -- which finally causes me to give up on it in disgust. (Which is what happened, and I did, the first few times I tried to get it to run at all.) -Kyle H On Feb 24, 2007, at 3:42 AM, Eugen Leitl wrote: > On Sat, Feb 24, 2007 at 03:10:04AM -0800, Kyle Hamilton wrote: > >> If someone can put together a VMware virtual appliance for it, it >> would work much better I should think. > > They're convienient, but the overhead for a VMware guest or Xen > is rather high. I prefer to run things in leaner virtual environments, > such as the Linux VServer. > > -- > Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org > ______________________________________________________________ > ICBM: 48.07100, 11.36820 http://www.ativel.com > 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE |
Kyle Hamilton wrote:
> Not everyone has the luxury of a VServer-based virtualization > environment. Also, not everyone has the time necessary to learn how > to administer one. > > If you want to continue with Croquet being a high-cost-for-entry > development environment, please be aware that you will merely cause > more and more people to be frustrated by the difficulty and veer away > from the promise that Croquet has to offer. If you would like to > make Croquet more easily accessible, please be aware that consumer- > level/non-OS-dependent virtualization is a low-cost/low-barrier means > to do so. > > If I can download a virtual machine that runs on Linux, configures > itself, and runs a headless server so that I can see how the client > interacts with it without having to have a separate machine, I'm more > likely to be wowed by it than by something I have to wrestle with > just to get only a red screen out of -- which finally causes me to > give up on it in disgust. (Which is what happened, and I did, the > first few times I tried to get it to run at all.) > > -Kyle H > As far as I know, all of the issues you raise are on the current development agenda. In particular, the Red Screen of Death on some unix/Linux systems seems to have been definatively solved, and will be available in a new release quite soon. -- Ed Boyce > On Feb 24, 2007, at 3:42 AM, Eugen Leitl wrote: > >> On Sat, Feb 24, 2007 at 03:10:04AM -0800, Kyle Hamilton wrote: >> >>> If someone can put together a VMware virtual appliance for it, it >>> would work much better I should think. >> >> They're convienient, but the overhead for a VMware guest or Xen >> is rather high. I prefer to run things in leaner virtual environments, >> such as the Linux VServer. >> >> -- >> Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org >> ______________________________________________________________ >> ICBM: 48.07100, 11.36820 http://www.ativel.com >> 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE > |
In reply to this post by Hussayn Dabbous-2
Hi,
As promised at the Intermetaverse meeting here is a demo of interoperatiblity between Croquet and Second Life. It leverages a central idea behind Croquet - replicated messaging. I'll comment specifically on your message tomorrow. Thanks to all who attended the meeting it was very encouraging! IM Khemical Stonecutter in world for more info. On 2/23/07, Hussayn Dabbous <[hidden email]> wrote:
hi; |
In reply to this post by Hussayn Dabbous-2
> If we can get MANY individual setups interconnected, we could very quickly get really > huge. And we will see whether the concepts can support that... > If we had a popular download button [download your grid point server] which contains > everything (including the installation shell script) admins could fire up their grid > point in less than 5 minutes, that would be the beginning of a shockwave I'm very interested in seeing a "croquet appliance" created that would allow a server to be brought up headlessly and configured with minimal expertise. I'm willing to donate some time & energy to making this happen using the rPath Appliance Platform, which I know very well indeed. However, I need a collaborator who knows Croquet very well indeed. I note that someone with handle 'pshouse' started a Croquet appliance project on rBuilder Online recently (http://www.rpath.org/rbuilder/project/croquet/) I can help whoever that is. Regards Brett |
Free forum by Nabble | Edit this page |