== How to install html5lib ==

=== Mac OS X 10.4 ===

Step 1: The default installed python is too old (2.3), so you need to install a newer version. Download and install [http://python.org/download/ Python 2.5.1 for Macintosh OS X]

Step 2: Download [http://html5lib.googlecode.com/files/html5lib-0.9.zip html5lib-0.9.zip]. Extract the zip file by double-clicking on it in the Finder.

Step 3: Open Terminal.app and type:  "c", "d" and a space. Drag the html5lib-0.9 folder into the Terminal.app window, so it prints the path name.  Type return.
Then type "/usr/local/bin/python setup.py install" and return.

=== Debian etch ===

Step 1: The default installed python is too old (2.4), so you need to install a newer version. Run "apt-get install python2.5-dev"

Step 2: Download [http://html5lib.googlecode.com/files/html5lib-0.9.zip html5lib-0.9.zip].

Step 3: Extract the zip by running "unzip html5lib-0.9.zip"

Step 4: Install it by running "/usr/bin/python2.5 setup.py install"

=== Why not use JTidy? ===

The initial version of Jabhts used [http://jtidy.sourceforge.net/ JTidy], but 

1) JTidy is buggy, e.g. removes empty <div> elements even when told to keep empty paragraphs.

2) JTidy is not maintained. The last release was made in August 2000.

As JTidy doesn't use the [http://www.whatwg.org/specs/web-apps/current-work/#parsing contemporary rules] for parsing html, and hasn't been updated for 7 years, it was dropped even though installing the python version of the html5lib takes much more work.