[Doc-SIG] Simplified table syntax proposals (reStructuredText )

Moore, Paul Paul.Moore@atosorigin.com
Thu, 27 Jun 2002 12:35:45 +0100


From: David Goodger [mailto:goodger@users.sourceforge.net]
> I've added some new, simplified but limited, alternative table syntax
> proposals (http://docutils.sf.net/spec/rst/problems.html#tables,
> alternatives 3, 4, and 5), incorporating an idea from Simon Hefti.
> There's a better example of alternative 3 in the notes.txt file
> (http://docutils.sf.net/spec/notes.html#bugs).
> 
> Could people please take a look at these alternatives and tell me what
> they think?  Thank you.

[[Disclaimer: I don't use tables much in raw text, so I don't have much of
an opinion from a writer's point of view. This is basically a view in the
context of being able to read the markup in raw form, which I feel is
important...]]

I find option (3) almost totally unreadable. Option (5) is OK, but the
inability to have blank cells in column 1 makes tables like this::

    +-----+-----+
    |     |  A  |
    +-----+-----+
    |  1  |  A1 |
    +-----+-----+

unrepresentable. And I suspect that such tables are fairly common
(spreadsheet type layouts).

Option (4) is the best of the 3, from that point of view. But I don't like
the (mis?) use of the bullet list notation.

I'd stick with the current approach. It's verbose, but very clear. The need
to pad out columns to type the | characters is a (mild) nuisance. But for
that, option (2) is fine. I feel that minimalism beyond that point is
useless - it hampers readability and expressiveness, for no clear gain in
ease of use.

[If reST ever grows a way of declaring options for a document, maybe having
a "table-format=boxed/minimal" option to choose between table formats (1)
and (2) would be worthwhile, but I imagine tghat the price is not worth the
gain. No, stick with things as they are.]

Paul.