I think it is worthwhile to explore to what extent Autoformat as you type is useable and useful for documents that are pure unicode text or unicode programming source code.

(I am assuming that the file is saved in a text format with a unicode option such as utf-8 selected. Source code needs the file extension changing from .txt outside Word.)

Effects that are implemented with unicode characters (including the common ASCII characters) will still work in a text environment.

The effects that are still available include:

• matching parentheses.
• smart quotes
• fractions
• bulleted lists
• numbered lists
• extended dashes
• smileys
• first line indents

At the present time, many compilers accept arbitrary unicode characters in comments and string literals. There are certain effects very likely to crop up in non-literal source code that most likely one definitely doesn't want to obtain by accident:

1. smart quotes
2. extended dashes
3. fractions
4. smileys

(1) and (2) are particularly insideous, because one may not notice their presence, and one may spend many frustrating hours debugging before realising the cause!

One can turn off these options, or have a vb-macro turn them on and off. Or one can have a cleanup vb-macro that recognises non-comment non-literal text and replaces the offending characters by the more sedate alternatives, and flags anything else non-ASCII.

Note that where permitted, one does want smart quotes and extended dashes in string literals.

There are plentry of opportunities to use autocorrect and building blocks as a (non-parameterized) text substitution processor (macro processor - definition sense) for executable source code as well as for text.