Jeremias Maerki wrote:
Patrick,
I have the suspicion that the auto table layout won't fill you up. With
the new design, basic support for auto table layout shouldn't be a very
big deal. There are some details, however, that could make this a much
bigger fish:
I did suspect it would not be enough.
Determining ideal column widths based on the inline elment lists for the
whole table should be relatively easy. You would probably have that
within two or three weeks after a, say, four-week get-to-know-FOP-better
phase. It gets more difficult if the column widths are to be
redetermined for each page (if that's requested). FOP cannot currently
do that because it determines page breaks on a number of pages at the
same time (total fit algorithm instead of first fit (Knuth terminology)).
This would cause extensive changes to the whole layout engine but enable
other features that are currently not possible, for example changing
available IPD between pages.
Sounds interesting, and it would be a good challenge for me to tackle.
The four-week get-to-know-FOP-better period you suggest sounds very
reasonable to me.
Another thing that just crossed my mind would be a total refactoring of
the images package. I'm itching to do that for a long time now but
probably won't have time for it in the near furture. I've written down
some thoughts about that [1] and can elaborate if necessary.
[1] http://wiki.apache.org/xmlgraphics-fop/ImageSupport
Sounds good too. I'll have to trust you about what kind of workload this
would all add up too. I'd like to be as realistic as possible since it
is important to meet the initial objectives of the proposal. In any case
I am motivated to actively contribute beyond the Google SoC.
Patrick
