TODO

 - dots issue
  - AdobeXML*.pdf
 - font issue 
  - cjkmix2*.pdf
 - space issue
  - python简明*.pdf
 - smarter space-as-offset
 - show hints about possible useful parameters
 - optimization levels
 - don't dump image when it is empty
 - spinning icon for dynamic page loading
 - fix negative word-spacing for IE

== Future: ==

Too difficult/complicated to implement:
 - integrate splash/cairo
 - naive image/drawing (SVG?)
 - type 3 fonts (convert to SVG fonts?)
 - reflowable text/combine lines/unwrapping
 - multi-thread

Not enough motivation/Lazy
 - commandline argument auto-completion
 - use absolute positioning for long whitespace
 - detect duplicate base fonts when embedding
 - disable selection if we know unicode is wrong
 - check if we can add information to the font, and let browsers show ligatures automatically
 - position history stack (popstate)
 - draw non-orthogonal lines with CSS
 - rename single-html -> embed-font/image/css ...
 - precise link destination: zoom
 - multiple charcode mapped to a same glyph
 - minimum line width of css drawing
 - ajax in pdf2htmlEX for separated pages
 - separate classes for annotations (such that we don't have to hide all css drawings for printing)
 - Widget Annoataion
 - handle large negative letter space
 - different optimization levels