Have detached proofs cross-reference their theorems
Have a detached proof function like an xref.
<proof ref="a-big-theorem" text="type-global">...
Somewhere early in the presentation of the proof, have a knowl/link to the theorem itself.
We could have the theorem link to multiple proofs by "discovering" the ref that point to it.
+1. In proof/logic text I frequently want to give multiple proofs (perhaps even a couple of incorrect ones too) of a given statement, so the ref is important.
There could be an option, maybe the default, for a cross-ref's proof to appear as
Proof of Theorem 3.14
Since such proofs are not children of their theorem-like, they should probably be numbered, e.g.
Proof 3.17 (of Theorem 3.14)
where Theorem 3.14 is a knowl, appearing when @ref is defined to be its @xml:id.
Good progress today on this - could be pretty efficient.
Exactly, but then we need to internationalize "of". And there are two knowls in close proximity?
Been leaning towards
Proof (Theorem 3.14)
And still debating with myself...
But yes, @text is not necessary and will be a reasonable default.
On March 9, 2022 10:12:14 AM PST, "David W. Farmer" @.***> wrote:
There could be an option, maybe the default, for a cross-ref's proof to appear as
Proof of Theorem 3.14
It is common and reasonable (maybe universal and reasonable) to not number proofs. If you go to the effort to make proofs numberable, I suggest there should be an option to not number them, and that should be the default.