I've been improving the .doc import of KWord over the last couple of months (besides working on tables)
I must say it has been quite a dramatic improvement. From having each paragraph being loaded onto each own page with the formatting all wrong, no tables, page size wrong, and many other bugs all over, I'm pleased to say that I'll this weekend give the .doc import a real world test, as I'll be conducting a 2,5 hour lecture using KWord to go through a 40 page document with tables, images inside tables, pagebreaks and lots of formatting.
I'm quite excited as this will be the first indication to me that KWord is getting ready for real use. We are not there yet for everyone, but I certainly intend to be using KWord as more than a .doc viewer in the future. That future still requires some hard work which I'm glad to be doing as my full time day job employed by KO GmbH
Syndicate Blogs
why are the specs so complicated
Here is a very well written blog about the M$ office formats, hope it helps although it is mainly about Excel: http://www.joelonsoftware.com/items/2008/02/19.html
answering my own comment
in this article, Joel talks about 1000s of man-years needed to produce an office importer. I guess 1000 man-years might be enough (depending on the team) as the first man-years of development Office were done on slow computers and without productivity gain e.g. from GUI tools. Maybe, if your goals is to be just as good as OpenOffice, already 100 man-years are enough to write the importer. And, as we know how it is with the man-years, maybe 50 years of your time would already be enough to make the import as good as the one from OpenOffice.
However, this is just guessing around. It would be good to get a figure like "OpenOffice used 500 man years for the Office import", but as far as I know, I am the only person documenting the time he needed for doing code changes.
All in all, I do not know if you have 50 years left in your life at all. I would URGENTLY recommend looking at OpenOffice's import code and learn as much as possible. Remember, saying "the code is a mess, I can do it better and I cannot even learn from the code" is normally what bad programmers do.
switch from Open Office ?
That's great news. I would be really happy to be able to switch from OpenOffice. But does your work include work on the export filter ?