>> move notes from TODO into the man page and "FAQ"

README.md

@@ -20,8 +20,9 @@ characters, at least if you don't want to manually write a very complex
 regular expression every time you encounter this problem.
 
 Owl will hunt down any files with these issues and rename them as 
-appropriate - see `man.md` for usage.
+appropriate - see `man.md` for usage and other important information.
 
+See [Infrequently Asked Questions](./iaq.md) for unimportant information.
 
 ---
 

iaq.md

@@ -0,0 +1,40 @@
+# Infrequently Asked Questions
+No one but has asked these questions, but I'm giving answers anyway, mainly 
+so I don't forget them in future.
+
+## Why didn't you use Punycode?
+[Punycode](https://en.wikipedia.org/wiki/Punycode) is a web standard for 
+converting general Unicode text into ASCII, for use in domain names, and for 
+nothing else. As the Wikipedia page says, it maps "München" to "Mnchen-3ya", 
+which I personally find harder to understand than "M_Ufc_nchen". It also 
+would add a dash to the end of pure 
+[ASCII](https://en.wikipedia.org/wiki/ASCII) text, so "London" would become 
+"London-", and its output does not handle characters that are bad for 
+shells, like "?".
+
+## Why didn't you use `base64/32/16`?
+These are internet encoding standards used to convert any binary data to 
+ASCII text, which sounds perfect for the job. However, their outputs are not 
+human readable, and Owl is designed so that you will at least know roughly 
+what the file was originally called.
+See <https://www.rfc-editor.org/rfc/rfc4648.html#section-10> for examples.
+
+## Why not just `mmv`? Or `zsh`?
+[mmv](https://github.com/rrthomas/mmv) is command line tool to do a kind of 
+"glob search and replace" with filenames. [zsh](https://zsh.sourceforge.io/) 
+is a shell, Bash or the PowerShell, but it has a tool to do a similar thing 
+called `zmv`. These require writing a regular expression of glob pattern to 
+represent the changes you wish to make, which is fiddly work when what you 
+want is to exclude potentially thousands of characters, not to mention 
+potential issues with hidden files (a.k.a. "dot files") and shell globbing.
+
+### Other alternatives
+- <https://github.com/DDR0/fuseblk-filename-fixer>: This actually does 
+  pretty much what Owl does, but is more specialised to its author's needs.
+
+## Tree
+This is used for testing. It should be available on any Linux system from 
+your distribution's repositories. If not, the source code can be downloaded 
+from the GitLab repository at 
+<https://gitlab.com/OldManProgrammer/unix-tree> or from the home page of the 
+project at <https://oldmanprogrammer.net/source.php?dir=projects/tree>.

man.md

@@ -33,3 +33,29 @@ Show version information.
 
 ### -h, \-\-help
 Show a small help message.
+
+
+## Edge Cases & Other Tidbits
+## What doesn't FAT allow?
+Mainly `*<>\|/:?'`
+
+### What about whitespace?
+Some of these (like tab, line feed, and carriage return) are not valid under 
+FAT and will be removed.
+
+If a sequence of invalid characters surrounded by spaces is in one of the 
+file names, and the `remove` strategy is being used, then you will be left 
+with multiple whitespace characters in a row. For example,
+```
+Really questionable ?? filename
+```
+will become
+```
+Really questionable  filename
+```
+with a double space before `filename`.
+
+### What about invalid UTF-8/UTF-16?
+This is removed and replaced with "\_INVALID_". This replacement is done 
+before anything else.
+