I would be in the revolutionary camp as well.
It seems to me that many issues and a whole host of limitations are spawned directly from trying to remain fully LLLP compliant and I seriously doubt that many of LL's choices (made ~15 years ago in the initial design phase) would be repeated if you were sitting down to design a system of this nature, today, from scratch, using the technologies of 2014 (not to mention all the things anticipated in the pipeline). In many respects Opensimulator development has boiled down to only two things: (1) maintaining "currency" with whatever feature additions LL is offering; and (2) new unique features that can only be achieved if they fall within the framework of LLLP. I can't begin to enumerate the number of excellent ideas I've heard that are instantly shot down only because "that would require a viewer change and we can't do that". Even using existing architecture, many things could be added if there was a dedicated Opensim viewer that could take those steps forward hand in hand with the simulator. I would even suggest that if a revolutionary approach were taken, support for LLLP and existing content ought to placed relatively low on the list of design goals to avoid hamstringing the whole process -- LL themselves obviously agree with that sentiment because they've stated that SL2 is unlikely to be backwards compatible other than perhaps in a very, very rudimentary way. I'd follow that lead and design something that meets the needs and even dreams of the people who will be the anticipated user base of the future; and then if it's possible to somehow bring in existing content, super. Honestly, I think any product that satisfies the "needs and dreams" goal will probably make 90% of existing content obsolete anyway...who would want to use "old crappy stuff" instead of taking advantage of the shiny new toys? A radical departure and loss of existing content is certainly going to lose some users who simply can't bear the thought of starting all over again -- or who decide on another alternate platform when forced to make that change -- but has just as much if not more potential to expand the user base if done correctly. All of which is lovely; but the question that arises in my mind almost immediately, regardless of which avenue is taken, is where the necessary manpower is going to come from? The number of people actively working on Opensimulator isn't even enough to address the existing bugs or keep up with the latest LL additions, let alone take on a project of the sort of magnitude a new/revamped viewer and simulator would require.
_______________________________________________ Opensim-dev mailing list [email protected] http://opensimulator.org/cgi-bin/mailman/listinfo/opensim-dev
