Should we allow disabling a witness

[swcurran]

The following questions came up in the review of PR #136. Let’s discuss and resolve them either by not adding them or adding a new PR.

• @brianorwhatever : I don't think we say anything about whether witness can be disabled after it's been enabled or is this forbidden?
• @brianorwhatever : Confirm the wording around “current or later” proofs of versions and whether it is in the right place.
• @PatStLouis : Wouldn't just signing the versionId be sufficient since the versionId is a sort of hash of the entry? Meaning if the entry validation passes, a proof over the versionId only would be sufficient?
• @PatStLouis : Does it make sense to instead only keep the proofs about the two most recent log update entries?

[swcurran]

RE: Disable Witnesses. I think we can add a line that allows for including a ”witness” : [] as valid. That entry would have to be approved by the active witnesses, and thereafter there would be no active witnesses.

[swcurran]

This is dumb to have these as one issue. Will move the other issues to new issues. Doh!