Some items that needs to be done:

Urgent fixes:

GUI/framework:

Parsing:
 - do not reparse the whole document when a <? or <script > or similar thing is
   added. Now the whole reparse is triggered if the parsed (invalid) area has a
   special block without end, like <? without ?> or <script> without </script>.
   This will speed up Quanta, but is quite complicated. If the current solution is good
   for users, don't bother with it.
 - do not reparse the whole special area when something is changed inside.
   Reparse only the changed nodes. This will speed up Quanta, but I saw that it can be
   very complicated.
 - do not reparse the whole file when a text is moved inside the document
   (cut&paste in one step, like when doing d&d). I don't have any idea how to do
   it... This will speed up Quanta.
 - get rid of the global baseNode
 - write a DESIGN file about the parser (and add more documentation to the header files)
 - write a parser regression tester

Structure tree:
 - do not delete and rebuild the structure tree when it's not necessary. Delete and rebuild
 only the changed nodes. This will speed up Quanta.

Plugins:
  - Kallery
  - config dialog rework

Misc:
 - WYSIWYG: table editing? layouting?
 - remove unneeded header file dependencies (work in progress)

Bugs discovered:
 - autoclosing closes also single (unknown) tags
 - autoclosing does not always closes a tag
 - attribute tree is updated and redrew on every keypress in case of PHP

DTEP:
 - in feature try to use the WEML for storing the DTEP tags
 - better document type recognition, namespace recognition (<MyNS:rootElement xmlns:MyNS="http://www.example.com">)