--nofuzzies option
  use original rather than fuzzies when converting back from po files

--update option
  make recursive methods only run if input is newer than output
--force option
  only overwrite files if this is given (work out which tools its applicable to)

detect whether producing a .po or a .pot (make -P / --pot unneccessary)

moz2po
  make strings fuzzy if original == translation

headers
  charset=UTF-8

copyright
  need to insert the standard copyright text unitialised

csv2po
  * change order of columns: original, translation, other stuff
  * It does not currently reflow text to create a newline around literal \n
  making it difficult to check files
  * allow a --no-wrap option so that we remove PO line breaks that mess with 
  translators segmentation tools.
  * We do not handle languages with N != 2 plural forms

oo2po / po2oo
  fix problem with OOO_VENDOR and \" quoting in offmgr.po (see help - about)
  make oo output order the same as in input file...
  ? remove the comment before msgid, treat all msgids as duplicated, thus inserting _: <comment>\n in msgid

pot2po discussion with pavel:
  pot2po should be as simple as using pot as one arg and po as the second ;-)
  pot2po pot po -t po warns me for every file: pot2po: warning: writing to temporary output...
  pot2po pot po -t po does not merge existing translations from old POs that have more then one comment and thus are duplicated in the new POs

docs
  not sure where else to put this!
    docs are installed in /usr/lib/python2.3/site-packages/translate/doc
    should be /usr/share/docs/translate-toolkit-0.8
    Not sure what's needed on Windows - non critical

po2dtd
  when reporting ampersand check failures should tell you the file not just the
    name of the bad PO file not just the command line output name

po2xliff
  * Handle PO plural conversion