Character subsets of words are matched as acronyms

The plugin is designed so that <acronym>[<id>] can be matched, so it's not so straightforward to avoid matching WFMD. We could avoid these matches in the case that an ID URL is not specified for the acronym, however we'd still get false matches in some situations.

For example, with the following definition,

||RFC ||Request For Comment $1 ||http://www.ietf.org/rfc.html ||http://www.ietf.org/rfc/rfc$1.txt||

we'd still match RFCD, when we only intend to match trailing numerics such as RFC2317.

I can think of a couple of options:

• Make the acronym definition syntax more sophisticated so that, for example, only trailing numeric characters would be matched in the example provided above (replace $1 with %d).
• Rather than matching RFC 2317, create a wiki markup syntax such as RFC(2317). This could be transformed to HTML as RFC 2317, as is currently done. The major issue I see with this is that it is not backward compatible, however we could do the work on a new branch (e.g. 0.12), or tag the current version before moving forward.

Let me know what you think.