Of all the features in S3D, supports give me the most issues by far. There are a few improvements and new features that I think might help alleviate some of the problems associated with them.
1. When placing new manual supports after auto-generating it would be very helpful if you could display all existing ones as semi-transparent. This would help in finding voids or blank spots that might need to be filled with an extra support or two. The same ability would also assist in finding ones you want to manually remove.
2. Similar to the first request, it would be great to be able to add manual supports when Cross Section view is enabled. Currently any faces that are being hidden in the cross section clipping don't seem to be valid attachment points and you're not able to place a support there. Many times I need to use cross section to see complex shapes hidden by other geo and supports that might need supplemental pillars.
3. A mode to disable supports display in the view without the need to remove them would also be very useful, or as mentioned before a see through mode where they draw semi-transparent.
4. An optional flag you can set for "helper discs" aka Mouse Ears that prevent the "Lower Vertical Separation Layers" setting being applied. This will force supports to have zero separation from them, thus better adhesion and less chance of failure.
5. Another option similar to above, but instead to force the RAFT separation distance to NOT be applied to support structures. In my opinion this is really only useful for easier separation from the PART itself, but to have the supports not adhere to the raft does more harm than good.
6. Better detection of isolated and unsupported "dense" support layers when they change from the sparse to dense values. I find that quite often sections of the dense support layers are simply floating in space due to the change in draw pattern.
7. Ok, this is a BIG one. I've posted about it before and I know many others did not agree that it was needed. However after several more months of using S3D I can attest to the fact when doing lots of portrait bust prints it would be a LIFE SAVER and get rid of many of the artifacts and print failures I'm running into. What I'm talking about is a method to have the BOTTOM of supports contour to the shape of the surface below, similar to how the upper part of supports do now. The current way supports are generated create pillars that overhang on steep surfaces such as the chest or back of neck of a human bust. What ends up resulting when sliced is a support section that attaches only on the very edge and then the rest of the support juts out into open space thus falls apart. The only way I've found around this is to use VERY small support pillar resolution (0.5-1mm), but these are extremely tedious to manually add and edit. An alternate solution would be a feature that would take your current supports and "subdivide" them into a smaller resolution after placement. In other words, you would add them as 4.0mm for example, then run the subdivision process which would not only put smaller pillars in place of the existing ones, it would also re-adapt the attachment points so you get more accurate contours on both top and bottom surfaces. Make sense? This would be my NUMBER ONE wish for S3D right now since it would make a huge difference in my print quality and reliability.