(Ray Saintonge saintonge@telus.net):
To me dumbing down means depriving people of their own challenges in life, and doing things for them that they should be doing for themselves. Mothers do this all the time when they compulsively pick up their children's things; the result is children who never learn to pick for themselves - much to the irritation of their eventual spouses. The educational component of Wikipedia is not just about content and the process of making that content satisfy NPOV. The same process can also apply to article stucture and markup. Need the Wiki vs. HTML dynamic be any different than the one about American vs. British English.
That's a good point. We probably don't want to limit the content model so much that people can't learn and grow to think in terms of document structure and presentation as well as content. So if putting in those more advanced structural elements is a little difficult, that's not such a bad thing. But disabling them entirely in the name of simplicity probably is.
Here's my main concern: I don't expect a newbie encountering a complex table to be able to edit the table without taking some time to learn what might be a complex syntax. But I /do/ want a newbie who sees a table with a misspelled word or an out-of-date statistic in it to be able to click edit, find the text with reasonable ease, and fix the text.
That's why including the table from another namespace is a total non-starter for me. When the newbie who clicks edit opens the page, he may see something simpler, but he still won't see the word he wanted to change, and he won't know how to get there, so we haven't solved the problem at all.
Here's another proposal for how we might handle more complex markup like nested tables:
In my current vision, this:
{{xxx text.... * list... text... }}
already creates a DIV block-level element in order to apply the style "xxx" to it, and
{{xxx Inline text...}}
similarly creates a SPAN. Let's (1) allow omitting the style tag altogether, so we can just create an anonymous DIV or SPAN, (2) make the parser keep a stack of open DIVs and SPANs, and allow elements to nest inside them, and (3) make tables usable at phrase-level like images (which they are in HTML anyway).
That way, the nested table looks like this:
||| === Title === | head1 | head2 | head3 | {{ || ==== Inner table ==== | cell | cell }} | cell | cell | cell
This would also allow you do things like putting paragraphs within list items, as well as nasty things like putting a table inside a list item, which gives me pause. Anyway, it's just another possibility. It does have the benefits that (1) all the text is visible as text, with no alphabetic markup to obscure it (with the exception of any style tags that might be added), and (2) the blank line terminates the structure, and line breaks terminate rows, so if a user screws up the effects will be limited.