> It will be wonderful if ucs-insert shows the characters 
> visually in the completion buffer.

FWIW, Icicles has had that feature for quite a while.

The attached screenshot shows also what happens when cycling:
*Completions* mode line shows the char with its code point.

,----
| icicle-read-char-by-name is a compiled Lisp function.
| 
| (icicle-read-char-by-name PROMPT &optional NAMES)
| 
| Read a character by its Unicode name or hex number string.
| Display PROMPT and read a string that represents a character by its
| Unicode property `name' or `old-name'.  Return the char as a number.
| 
| You can use completion against the Unicode name of the character.
| 
| In Icicle mode:
| 
| * The character itself is displayed next to its name, even though it
|   is not part of the completion candidate.  WYSIWYG.
| 
| * When you cycle among candidates, the current character and its
|   Unicode code point are shown in the mode line (provided user option
|   `icicle-help-in-mode-line-delay' is greater than zero.)
| 
| If you use a dedicated `*Completions*' frame, then the font used in
| `*Completions*' is the same as the frame from which you invoked
| completion.
| 
| If you use library `doremi-frm.el' then you can increase the font size
| for `*Completions*' dynamically using `C-x -'.
| 
| As an alternative to completing the Unicode name, you can input a
| number for the Unicode code point: a hexidecimal number or a number in
| hash notation: #o21430 for octal, #x2318 for hex, or #10r8984 for
| decimal.
| 
| Non-nil optional arg NAMES is an alist of names to use in place of the
| value returned by `icicle-ucs-names'.  It must have the same form as
| such a return value: (CHAR-NAME . CHAR-CODE).
`----