For now, I did not yet decide which route I will take, but I do know which route I would prefer, and it would be staying with the Snowglobe v1.5 code base and keeping up the backporting work so to get the Cool VL Viewer code closer and closer to v2 over time, with the notable exception of anything UI-related.
The reasons are mainly that:
- The work necessary to bring the v2 UI back to an even remotely acceptable level of usability is HUGE.
- The work necessary to make v2 OpenSim-compatible is quite significant.
On the other hand, some forthcoming v2 features may simply involve too much code change to be backported to the SG v1.5 code base in a reasonable amount of time. Then, and only then, I would have no other choice than to switch to the v2 code base (yet still probably maintaining the Cool VL Viewer v1.25, at least till I could make v2 fully OpenSim-compatible), but by then I hope that other TPV teams will have already done some good work on the v2 UI, work that I could reuse (remember that being alone to develop the Cool VL Viewer, and even while I'm a rather proficient programmer, I can't really dedicate as much manpower as Imprudence or Phoenix teams, for example).
For now, the most complex forthcoming feature is Mesh support... I did not yet try to backport it, since LL did not yet set a date for its implementation on the main grid (also what will happen in Q1 2011 is already known
and doesn't involve anything Mesh-related), but from the (quick) look I had at the code, Mesh support might
well be portable to SG 1.5 with less work than what would involve reworking the v2 UI... If I can manage to do this backport, then I see no issue to keep the SG v1.5 code base for at least one more full year.