• Disallow the characters that confluence does not allow (try adding apage with a comma in it, and see the list in the error message - where is this info kept can we get it dynamically so if it changes we're reflecting the given Confluence's requirements)
• While still converting the page, and keeping links accurate

So....
I'm thinking a two step process
From within the ConverterEngine, we always add two converters to the end of the converter string list.
These converter do the following things:
1) Make a list of pagenames that are illegal
2) For each illegal pagename, create a substitute that is legal
3) Change the illegal pagename to the legal substitute
4) Look through every page for links to the illegal page name and change the link to the legal substitute.

(Step 4 requires that steps 1-3 have already been accomplished and saved as state (maybe a hashtable?), therefore step 4 is the second step in the "two-step" process. Also, therefore, need 2 converters. First converter does steps 1-3. Second converter does step 4.