I need some help moving from the concept phase to the requirements gathering phase on this:
Imagine I want to extend existing browsers (FireFox primarily, Safari, Chrome (multiple platforms would be nice)) to recognize a new filetype and display it 'correctly'. Sort of like the Adobe PDF plugin (which Safari no longer needs), but for something like (TeX) DVI or PostScript. This will not be embedded in HTML like a movie, or image, but the whole file, like a pdf.
Where would you start for each of these browsers? Would this be best done as an NPAPI plugin for all browsers or might there be a better route for specific browsers (like Apple's webkit plugins)? This is how SVG support kind of started with Batik and other plugins. Are there samples for making an Netscape Plugin API object built for and installable on Windows, Mac OS X and Linux?
What if the viewing of the source format was best translated to HTML (on the client side) rather than a fully rendered image (E.G. bolding and indenting for YAML, or syntax coloring for *.ini files or apache log files, perhaps even a language like go) would that change your answer as to how to approach this?