Date: Wed, 8 Mar 2006 23:14:02 -0800
From: User Bbearnes
Subject: Re: wiki markup
To: Alan Grow

All right, so I've got a multipart answer.

Part one is is I think the core wiki markup is a slighly modified version
of the original c2/usemod standard:

CamelCase and/or [[page to link to]]
[http://what.the.fuck.ever linktext]

---- -> horizontal rule

 initial space -> preformatted/monospace text

* -> bulleted list item
# -> numbered list item

= header 1 =
== header 2 ==
=== header 3 ===

::more indented/nested blockquote

-- -> &mdash, in some systems.

In general stuff follows plaintext e-mail "styles", with a notable

''<i>talics or <em>phasized''
'''<b>old or <strong>'''

Which is actually fucking stupid; it should clearly be *strong* and
_emphasized_.  I've been using a wiki at work that does this, since it's
part of Textile, which is one of its default available languages.

Which brings me to my second part - you could just use Textile, which does
some cool stuff.  I don't know that there's a library out there to
interpret it, but as far as I know it originates with a free licensed
blogging package called TextPattern, so there ought to be one.  Textile is
cool up to a point - for 90% of the markup you will ever do - then it
starts grating on my sensibilities.

Third part:  I know portability is always a concern, but I think it's
actually far less of one with decent wiki markup, since it's pretty much
the closest thing you're going to get with any real features to
human-readable plaintext.  It would be trivial to parse your wikitext out
into any necessary intermediary if you decided to migrate later - a lot
more trivial than it will be for the 6 years worth of HTML and embedded
Perl I've got sitting in a pile here.