The Tika plugin in http://issues.apache.org/jira/browse/NUTCH-766 is a first attempt at delegating the parsing to Tika instead of having to maintain the parser plugins in Nutch. This page will list the differences in coverage or functionality between the Tika plugin and the existing Nutch parsers. Tika also has more formats not covered by Nutch which are not described here and has a more generic capability of representing structured content which can be useful for HtmlParseFilters (which are currently limited to HTML content).
mp3: Nutch identifies several fields (Title, Album, Artist) whereas Tika knows only about Titles, the rest is stored as paragraphs.
Tika-app can also identify in an mp3 id3v1 and id3v2 tags like: album, artist, audioSampleRate, composer, genre, logcomment, releaseDate, trackNumber using the XMPDM interface
msexcel: comparable (+ Tika able to represent content in structured way as XHTML tables which can be useful for HTML parser plugins)
msword: Tika does not support word 95 other versions are comparable
rss: Tika identifies only the Mimetype but does nothing about the content
rtf: deactivated in Nutch for licensing reasons | works in Tika
swf : not yet covered in Tika (see https://issues.apache.org/jira/browse/TIKA-337)