It seems like everybody and his brother has a new format these
days. Yes, it'd be great if texinfo could write them all at will. But
we need a cleaner way to handle the multitude of formats. Otherwise,
maintaining all the different formats will be nigh unto impossible.
It seems to me that making a modular separation between the code that
parses Texinfo and the code that generates the output format should
not be very hard. How much has this already been done?
And that should be enough to make it fairly easy to support a variety
of output formats.
