Part Smoothing - Where Do We Go From Here?
2013-03-14, 3:24 (This post was last modified: 2013-03-14, 19:13 by Roland Melkert.)
2013-03-14, 3:24 (This post was last modified: 2013-03-14, 19:13 by Roland Melkert.)
Hi Y'all,
I'm sorry to start a third thread on part smoothing, but I wanted to step back and ask a process question; I haven't been involved in LDraw long enough to know how this gets sorted out.
There are a few ideas that have been thrown out on how to solve part smoothing problems, but they involve different groups of people and perhaps different standards bodies:
1. We might say that no change to the LDraw format or any parts or any change to authoring guidelines will happen, and programs need to take the most drastic steps while smoothing, e.g. code for the worst case.
2. We might intentionally modify the parts library, perhaps with mechanical transformations (E.g. run a program that splits T junctions over many parts). This would require some kind of approval.
3. We might intentionally introduce new syntax features into the LDraw file format (and then apply them to some parts) to simplify the process of smooth rendering.
Where do we go from here? Would it be useful for me to write up an RFC, just so we have a straw man?
I can continue coding assuming case 1, but such code will be non-optimal (e.g. we'll have to put in a part cache that we don't otherwise have just to make things performant) and such code would be obsolete if either 2 or 3 happen. If either 2 or 3 are on the table, it would be nice to plan for them.
cheers
Ben
I'm sorry to start a third thread on part smoothing, but I wanted to step back and ask a process question; I haven't been involved in LDraw long enough to know how this gets sorted out.
There are a few ideas that have been thrown out on how to solve part smoothing problems, but they involve different groups of people and perhaps different standards bodies:
1. We might say that no change to the LDraw format or any parts or any change to authoring guidelines will happen, and programs need to take the most drastic steps while smoothing, e.g. code for the worst case.
2. We might intentionally modify the parts library, perhaps with mechanical transformations (E.g. run a program that splits T junctions over many parts). This would require some kind of approval.
3. We might intentionally introduce new syntax features into the LDraw file format (and then apply them to some parts) to simplify the process of smooth rendering.
Where do we go from here? Would it be useful for me to write up an RFC, just so we have a straw man?
I can continue coding assuming case 1, but such code will be non-optimal (e.g. we'll have to put in a part cache that we don't otherwise have just to make things performant) and such code would be obsolete if either 2 or 3 happen. If either 2 or 3 are on the table, it would be nice to plan for them.
cheers
Ben