[Wikipedia-l] Wikification function proposal

[Magnus Manske]

A topic that was briefly discussed several times on this list in the past is
the automated "wikification" of pages. I think it is time to work on that,
is there is a need.

Some wish to ban all HTML code from the wiki articles, though I find this
rather extreme. Others prefer to use HTML as much as possible, especially if
copying from other (free) web sources.

IMHO, the best was is to use wiki tags where they are available, and leave
the complicated stuff (like tables) to HTML. A function that could help with
that could be invoked as follows:

1. After or prior to editing, as standard setting. This would enforce
wikification, but reduce the "editing freedom", thus probably not a good
idea.
2. After or prior to editing, as a user option. If you check that option on
your user page (default:off), wikification would take place when you edit an
article.
3. As a button on the edit page. Like "Preview", maybe "Wikify & Preview".

The functions that could be performed (and that I can think of) are:

1. Replacement of HTML tags by wiki tags (namely, "<b>" and "<i>", maybe
"<h1>" as well)
2. Eliminating links to redirect pages by changing the link in question to
the target of the redirect
3. Marking [broken links]] and [[more broken links]
4. Optimizing external links with too many [] (like
[[http://www.google.com]])
5. Optimizing [[link|links]] to [[link]]s
6. Removing [[Talk:thispage|Talk]] links (at the end of the page)

I am sure you can think of some more details that have always bugged you,
but that were (be themselves) not important enough to call for a function;)

We could use this for rather fancy things as well:

7. In the preview, mark words (or word combinations) that are currently
plain text, but that do exist as a topic in the database (to find pages one
could link to).
8. In the preview, mark all numbers that could be dates (more link
candidates).
9. Automatically put the title phrase in bold if it appears in the first
paragraph.

Thoughts? Comments? Threats? ;)