I maybe wrong, but I will say it anyways....
From what I have read, this is precisely the type
of applications that use XSL/XSLT. Have the content
in XML format. Use an appropriate style-sheet depending
on the client. Convert the output of the transform to HTML,
by setting the appropriate processing instructions.
This way you have ONLY ONE place you have
to maintain your content, as opposed to three
places/sub-classes in the method you mentioned.
Again, you probably have your reasons why to want to go
one way or the other. My knowledge is theoritical, anyways.
Thanks for listening.
regds.
- satya