A "face" is a collection of graphical attributes for displaying text: font family, foreground color, background color, optional underlining, and so on. Faces control how buffer text is displayed, and how some parts of the frame, such as the mode-line, are displayed.
For most purposes, you refer to a face in Lisp programs using its "face name". Each face name is meaningful for all frames, and by default it has the same meaning in all frames. But you can arrange to give a particular face name a special meaning in one frame if you wish.
We expose private value for free identity conversions when the value is nested in some covariant type, e.g. (symbols : Symbol.t list :> Value.t list) rather than List.map symbols ~f:Symbol.to_value.
eq t1 t2 = Value.eq (to_value t1) (to_value t2), i.e. eq checks whether the Emacs values underlying t1 and t2 are physically equal. This is different than phys_equal t1 t2, because we don't always wrap eq Emacs values in phys_equal OCaml values. I.e. phys_equal t1 t2 implies eq t1 t2, but not the converse.
The name of the "font foundry" in which the font family attribute is located; the wild-card characters `*' and `?' are allowed. (Info-goto-node "(elisp)Face Attributes")