[Wikipedia-l] Parsing

[Toby Bartels]

Lee Crocker wrote in part:

>Toby Bartels wrote:

>>We can do psychological testing to determine when simplicity
>>switches, but other people will switch it at slightly different
>>times. Best all around to allow both methods.

>This I have to disagree with totally.  Leaving in a mix of both is
>horrible.  Our markup here serves a different purpose than other
>markup languages. It /must/ be editable by novices.  Wikitext that
>produces a nice page, but can't be edited, is bad wikitext.  For an
>extreme example, see the article "Duesseldorf"--this is precisely the
>kind of abomination that allowing too much HTML produces.  It might
>be a fine article, but it will never be improved because it's
>impossible to edit.  That's fine for web page, but not for a wiki.

That example's not fair, of course.
First, it's not fine for a web page; the HTML is badly written anyway.
Next, disallowing HTML doesn't necessarily fix that;
see the first version, which had HTML that we didn't parse.
Now, that got fixed in a few minutes, so not parsing HTML discourages it.
In that vein, let me note that I agree that some HTML shouldn't be parsed.
So I grant your implicit claim that we allow too much HTML
(especially since Jan successfully put JavaScript inside a <nowiki>!),
but not your implicit claim (given the context) that <h1> is too much.

In fact, here would be a nice start at culling unallowable HTML
that Jan and I can both get behind (I hope): go to
http://www.w3.org/TR/1999/REC-html401-19991224/index/elements.html
and get rid of everything labelled "deprecated" (or not there at all).
This takes care of half of the problem with [[Duesseldorf]], which is <font>.
(I saw a talk or meta page somewhere that said that we might want <font>
in an article on a font itself in order to provide an example.
But somebody else replied quite rightly that an image would be superior.)

The other half of the problem is the table, which is completely inappropriate,
but we wouldn't be much better off if they used wiki syntax for it instead.
If we allow tables at all, and there are good reasons to do so,
then we'll have to deal with people using them inappropriately;
the computer nerds like us just need to rewrite them whenever we see them.

I don't for a moment think that we need to allow all HTML.