I understand you intend for application layers to be built on top of MXP. My point is without user-visible applications, reX doesn't exist, which makes it hard to argue that reX has to stop what it's doing and fund development of a complete MXP application stack.
If you believe strongly enough in MXP as the future HTTP, you should integrate a C++ client implementation for Naali. That would certainly get the ball rolling. Cheers, On Mon, Oct 26, 2009 at 2:40 PM, arkowitz <arkow...@gmail.com> wrote: > > Kripken - just what do you mean by "for what it does"? > > Ryan - the animated avatars, distributed assets, portable inventory, > and deployment tools you mention would fit nicely into the MXP model, > but would not be built directly into the "server". And I believe the > architecture enabled by MXP gives an advantage here by not expecting > everything to be built into a monolithic server. > > HTTP is not used merely for "transferring web documents", obviously. > It is used for much more than it was designed for, and indeed many > things it is not good for. But as a "one protocol" it has enabled a > fantastic culture of cooperation among server and client developers - > one group's browser or middleware works with another group's server or > service. You don't have the architectural silos that epitomize > virtual world platforms. > > Arkowitz > > > On Oct 22, 11:18 am, Ryan McDougall <sempu...@gmail.com> wrote: >> Is there an example you can point me to that demonstrates an MXP >> implementation that delivers a complete VW experience; from fully >> animated avatars, to distributed assets, portable inventory, >> deployment tools, etc? >> >> My point is given the point in time of our project, we have to deliver >> the best application platform in order to survive. The goal of having >> an "HTTP" is admirable, but people don't buy the protocol, they buy >> the application or service -- youtube or gmail. >> >> Moreover, you cannot compare the web and VW so directly. Sharing >> documents on the internet is an order of magnitude simpler problem. >> The reason we have so many VW protocols is that everyone thinks they >> can do something better, and there is no obvious right answer yet. >> This is a huge reason why Naali is intended to be a multi-protocol >> viewer: betting the house on one implementation is not wise. >> >> Never the less, the answer remains the same: we don't have resources >> to implement every conceivable protocol; we will get to it when we are >> able. If anyone would like to have it sooner than that, please stop by >> our IRC, and we can help you get it implemented in Naali. >> >> If anyone really things it makes better sense to implement MXP than LL >> protocols, there's a standing invitation to make that case. >> >> Cheers, >> >> >> >> On Thu, Oct 22, 2009 at 3:49 PM, arkowitz <arkow...@gmail.com> wrote: >> >> > The goal of supporting "all protocols" is admirable, but impossible. >> > The protocol used for virtual world communication will always >> > explicitly or implicitly affect the architecture of the system as a >> > whole. >> >> > MXP was created to allow virtual worlds to be architected properly. >> > One example of this is the way avatars are treated just like other >> > objects owned by participants and possessing awareness bounds. This >> > is an important principle and should affect the architecture of the >> > server. >> >> > Ultimately there should be one protocol. Think about it... how many >> > alternatives to HTTP are there? >> >> > Arkowitz >> >> > On Oct 19, 6:33 am, Ryan McDougall <sempu...@gmail.com> wrote: >> >> The ideal situation is that realXtend Naali supports all virtual world >> >> protocols, so yes one day we'd like to support MXP. Unfortunately, >> >> given how short of resources we are compared to the scope of the >> >> problem, we are limited by not having anyone come forward to implement >> >> MXP support within Naali. >> >> >> If you'd like to contribute to MXP support for Naali, please let me >> >> know. I've already asked Tommi Laukkanen to present a proposal about >> >> MXP, but understandably, he is also a very busy man. >> >> >> Cheers, >> >> >> On Mon, Oct 19, 2009 at 7:24 AM, mengzhehai china >> >> >> <mengzhe...@gmail.com> wrote: >> >> >> > Realxtend will use the MXP protocol in the future? >> >> > Opensim is using the MXP now. >> >> > Whether to build a service project using C++ or python If realxtend >> >> > use the MXP? >> >> > Many people can work for it ,the server will run on Linux/Unix better >> >> > than now . >> >> >> >http://www.bubblecloud.org/ > > > --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/realxtend http://www.realxtend.org -~----------~----~----~----~------~----~------~--~---
