+@node Q1.3.8, Q1.3.9, Q1.3.7, Introduction
+@unnumberedsubsec Q1.3.8: Does XEmacs support Unicode?
+
+Partially, as an external encoding for files, processes, and terminals.
+It does not yet support Unicode fonts @ref{Q1.3.9, Does XEmacs support
+Unicode Fonts?}
+
+To get Unicode support, you need a Mule-enabled XEmacs. Install
+Mule-UCS from packages in the usual way. Put
+
+(require 'un-define)
+(set-coding-priority-list '(utf-8))
+(set-coding-category-system 'utf-8 utf-8)
+
+Install standard national fonts (not Unicode fonts) for all
+character sets you use.
+
+Mule-UCS also supports 16-bit forms of Unicode (UTF-16). It does not
+support 31-bit forms of Unicode (UTF-32 or UCS-4).
+
+@node Q1.3.9, Q1.4.1, Q1.3.8, Introduction
+@unnumberedsubsec Q1.3.9: How does XEmacs display Unicode?
+
+Mule doesn't have a Unicode charset internally, so there's nothing to
+bind a Unicode registry to. It would not be straightforward to create,
+either, because Unicode is not ISO 2022-compatible. You'd have to
+translate it to multiple 96x96 pages.
+
+This means that Mule-UCS uses ordinary national fonts for display. This
+is not really a problem, except for those languages that use the Unified
+Han characters. The problem here is that Mule-UCS maps from Unicode
+code points to national character sets in a deterministic way. By
+default, this means that Japanese fonts are tried first, then Chinese,
+then Korean. To change the priority ordering, use the command
+`un-define-change-charset-order'.
+
+It also means you can't use Unicode fonts directly, at least not without
+extreme hackery. You can run -nw with (set-terminal-coding-system
+'utf-8) if you really want a Unicode font for some reason.
+
+Real Unicode support will be introduced in XEmacs 22.0.
+
+@node Q1.4.1, Q1.4.2, Q1.3.9, Introduction