On Nov 9, 2006, at 10:31, Jeremias Maerki wrote:
Hi Andrejus / Jeremias,
<snip />
Besides that I think it would be rather difficult to implement. I'd
rather not introduce tab stops in FOP. But I don't know what the
others
think.
Seems quite awkward to me as well, and only really feasible in case
of blocks of which you know for certain that they will not span
multiple lines. The line-breaking algorithm will become quite
complicated if it has to take into account such tab-stops (if I
esteem correctly).
Using tables, OTOH, would not necessarily be overkill, I think... :/
Starting from a fixed-layout (fixed column-widths, all 0.5in), the
only real effort would be to estimate how many columns the content of
each cell will need. Say at most 1in is needed for the first (but
also more than 0.5in), then its number-columns-spanned property
should get a value of 2. The second cell will then automatically get
column-number of 3, etc.
Apart from that tiny effort in the XSLT code, table-layout will take
care of all the rest without any intervention/ manipulation. It
remains a bit of guess-work, but in the end, I think using a table is
to be preferred over introducing such a non-standard FO.
Andrejus' observation that 'one row of such a table (is) independent
of another' seems to me not really true/accurate. That would be the
case if you needed different tab-stops for different lines. In the
example the tab-stops are all the same, so correspond to what one
could call virtual columns.
HTH!
Cheers,
Andreas
