These tables could include arbitrary lists of blocks in each cell.

+------------------------+------------+----------+----------+ | Header row, column 1 | Header 2 | Header 3 | Header 4 | +------------------------+------------+----------+----------+ | body row 1, column 1 | column 2 | column 3 | column 4 | | and more col 1 | | | and more | | and more | and more | and more | col4 | +------------------------+------------+----------+----------+ | body row 2 | Cells may span column | | +------------------------+------------+----------+----------+

Look at reStructuredText's table syntax as a paradigm.

Just for reference, this seems to have an overview: http://docutils.sourceforge.net/docs/user/rst/quickref.html#tables

And this has the complete specification: http://docutils.sourceforge.net/docs/ref/rst/restructuredtext.html#tables

I could see this as being useful for dealing with docs that contain these sorts of tables which you intend on converting to pandoc's markdown. Emacs has a mode for creating these sorts of tables. The MySQL client makes these kinds of tables. However, I think the table format that pandoc currently supports beats other formats in terms of elegance and ease of reading as plain text. (The only improvement I can see would be implementing the request I made in issue 156 http://code.google.com/p/pandoc/issues/detail?id=156 . )

So, I think it might be useful to have a pandoc option that would convert other table formats into the kind specified for pandoc's markdown. That is, something along the lines of, "filter this text file, and every time you see a grid-style table, turn it into a pandoc-markdown-style table".

Now that the rST reader has support for grid tables, it shouldn't be hard to add this to the markdown reader.

Grid tables have been added to markdown, as of commit cbac17cfd20ead56def539fdf0817b1297f36833.