Remove redundant traits (unless they have valuable proofs/references)

[StevenClontz]

See discussion in #55

[prabau]

Ideally traits that have valuable proofs/references but are also deduced could have the web site show both the explanatory text and the deduction. But that would be a longer term feature for the general pi-base infrastructure, maybe not so easy to do and not a high priority. Would have to decide if it's worthwhile or not, especially if there are only a few instances.

[StevenClontz]

So it's easy to model this: we just leave off the value: boolean in the YAML frontmatter. Then if the result is deduced, that's great, but if not we can show a ? with whatever discussion is in the file.

This is probably a good way to handle placeholders for results independent of ZFC, since pi-base is only equipped to handle one model of set theory at the moment.

[StevenClontz]

After discussion with @jamesdabbs we're going to just have the viewer expose redundancies, and human editors can decide if the redundancy should be removed. pi-base/viewer#73