Re: [mmox] MMOX: Strawman scope/goals/approach

The ITEF really doesn't do reference implementations. The bakeoff approach 
is sort of antithetical to
a singular reference implementation. The bakeoff approach says "You need 
several to prove 
interoperability" and you validate the spec, based on how well they 
interoperate, collectively, not
singurally. Now, of course, you sort out which ones don't meet the spec, 
and thus fail to interoperate,
but.. you also look at the set of interoperating implementatoins, ont a 
singular one. 

mmox-bounces@ietf.org wrote on 02/26/2009 09:04:45 PM:

> yes, but a reference implementation is not a work item on the group's 
> charter.
> 
> and personally... ugh.. i've seen waaay too many times when specific 
> implementation decisions for reference implementations were mistaken 
> for protocol requirements and pain ensued.
> 
> i even prefer the term "example code" instead of "reference 
> implementation" for this reason.
> 
> -cheers
> -meadhbh
> 
> On Feb 26, 2009, at 5:36 PM, Gareth Nelson wrote:
> 
> > A reference implementation is always a handy thing to have :)
> >
> > On Fri, Feb 27, 2009 at 1:32 AM, Meadhbh Hamrick (Infinity)
> > <infinity@lindenlab.com> wrote:
> >> i don't think we get to mandate what code people implement the 
> >> protocol in,
> >> but i would certainly add lisp, smalltalk and occam to the list.
> >>
> >> seriously thought... remember... we're talking about a wire- 
> >> protocol, not a
> >> library that implements the protocol.
> >>
> >> On Feb 26, 2009, at 2:31 PM, Jon Watte wrote:
> >>
> >>> Gareth Nelson wrote:
> >>>>
> >>>> any code should be produced in. My vote is for python and C.
> >>>>
> >>>
> >>> Let me vote for Lua and C++, and then the OpenSim people can vote 
> >>> for C#
> >>> and Java and we have all the bases covered :-)
> >>>
> >>>
> >>>> games are relevant (after all, MMORPGs for example are just huge
> >>>> virtual worlds with game rules thrown in), I fail to see how 
> >>>> networked
> >>>>
> >>>
> >>> I think that's a common mis-conception. Most (almost all) MMOGs do 
> >>> not
> >>> have much in the way of persistency, other than character 
> >>> possessions and
> >>> stats. They also have a very static world, that generally does not 
> >>> allow
> >>> users to move things around or place new things into the world. 
> >>> Finally,
> >>> they do not support any form of UGC, which I think is more or less a
> >>> requirement for a "real" virtual world. Lacking all three means 
> >>> that the
> >>> technical job of implementing and running the MMORPG is much 
> >>> easier, but it
> >>> also puts them on a lower rung of the expressiveness. A VW could 
> >>> express a
> >>> MMORPG (with significant overhead compared to the MMORPG itself), 
> >>> but a
> >>> MMORPG couldn't express a VW.
> >>>
> >>>> Practically everything people do online is "multiuser", and lots of
> >>>> things are "massively multiuser" - look at any big site such as
> >>>> wikipedia for example.
> >>>>
> >>>>
> >>>
> >>> "What is virtual worlds" has been a general question for a long 
> >>> time, but
> >>> I believe the "rough consensus" answer is that it's at least:
> >>>
> >>> Interactive
> >>> Persistent
> >>> Identity Based
> >>> Collaborative
> >>> Supports UGC
> >>> (controversial) Based on a 3D Simulation
> >>>
> >>> But, if we believe that we all need to agree on exactly what a 
> >>> virtual
> >>> world is before we can make progress, experience shows we won't be 
> >>> able to
> >>> make progress at all.
> >>>
> >>> Sincerely,
> >>>
> >>> jw
> >>>
> >>> _______________________________________________
> >>> mmox mailing list
> >>> mmox@ietf.org
> >>> https://www.ietf.org/mailman/listinfo/mmox
> >>
> >>
> 
> _______________________________________________
> mmox mailing list
> mmox@ietf.org
> https://www.ietf.org/mailman/listinfo/mmox