[Opensim-dev] Perl vs C# UGAI?

[Brian Wolfe]

I couldn't agree more.

My personal view of a project like OpenSim is that it is a reference
implementation of the OpenSim protocol(s). As such, it shouldn't concern
itself with containing every single implementation of the various
servers.  

I really like the idea of having a wiki page listing "OpemSim Protocol
based Projects" that end users can pick from. the opensimulator.org
website could focus on protocol specifics, and linking to specific
implementations with the core C# implementation downplayed as just a
reference implementation that may or may not be as stable or fast as 3rd
party implementations and utilities.

This would also make it possible for us to take the add-in modules like
the recent bad behaving interface demo module and keep them in a
seperate "official add-ons" repository that people can overlay on the
core report. How that overlay is acheived can be worked out later. ;)

I have to agree with Charles on the OSGrid concerns. I say leave it
using just the C# core reference implementation so it's a solid known
setup to test against. If we really want to semi-officially suport the
alternatives, then OSGrid can negotiate hosting of secondary testing
grids with the various implementations if the alternative creator
doesn't have a place to run their servers (I'm making an assumption of
resource availability here, which may or may not be possible).

On Thu, 2008-04-03 at 10:04 -0400, Sean Dague wrote:
> On Thu, Apr 03, 2008 at 02:27:07PM +0100, Justin Clark-Casey wrote:
> > I'm very much in favour of the idea of alternative implementations of 
> > the UGAI protocols.
> > 
> > As James, has suggested, in other circumstances it would be good to 
> > formally write down the protocol and advertise and discuss changes 
> > beforehand.  However, the problem is, as Michael and Sean say, is that 
> > it's still in a state of considerable flux (this is Alpha code!). Trying 
> > to formalize at this stage would considerably slow down development.
> > 
> > I have to agree with Michael that we shouldn't have alternative 
> > implementations in our own svn tree - it will cause considerable 
> > confusion as to what OpenSim officially supports and what it doesn't.  
> > And hypothetically, if Lulurun goes away for whatever reason 
> > (hypothetically!) the onus to maintain it as an 'official' alternative 
> > would fall on the OpenSim developers.  Whether the Perl alternative 
> > should really be the reference (not necessarily the best) implementation 
> > is another argument, I think.  I can see pros and cons for both.
> > 
> > Having said that, it sounds like Lulurun is willing to maintain the code 
> > to match changes in the protocol.  Even if the code doesn't live inside 
> > the OpenSim tree, I believe we could make an effort on an informal basis 
> > to advertise and discuss proposed protocol changes before the fact 
> > (unless the changes are very large, in which case things would have to 
> > be done post-facto).
> 
> Yes, we should have links to alternate implementations somewhere on the
> wiki.  Honestly, I suspect that we'll end up with UGAI in at least perl,
> python, ASP.NET, java, and ruby by the end of the year that people are
> maintaining.  That is a strength, not a weakness.
> 
>     -Sean
> 
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev