Synopsis - Issue Tracker

Right now comments are processed in multiple steps,
from filtering over parsing to formatting. This is
overly complex.
While the filtering has to be done in a separate
step (due to the fact that we support different
languages with distinct rules concerning what
constitutes a comment), there is no need to separate
parsing and formatting.
However, as different source files may use different
comment conventions, a processor is needed that
can annotate individual source files with the
markup style that is to be used.
Then formatters can provide facilities to parse
and format comments based on that annotation.

Most of this is now implemented. A 'translator'
generates 'doc' annotations from 'comments' 
annotations, and allows to set a markup attribute,
which is then used during formatting.

Some Python Enhancement Proposals seem to suggest
to embedd a '__docformat__' string into each
module, which makes passing a markup specifier
to the document extractor redundant.
Synopsis should honor such embedded metadata.