A
Andy Dingley
decided to do this with a wiki. The trouble is that it isn't
automated enough and it isn't smart enough. Sometimes you really do
need an application and not just a document.
A good wiki isn't just "a document" though. Speaking about MediaWiki
you have the categorization mechanism to annotate entries (pages),
templates to build "tags" to make this easy (a single template that
categorizes for you) and then extensions like DPL (dynamic page list)
that allow quite powerful querying on this. You're still constrained
by the mapping between "smallest unit of data" and "page" for input,
but you can build fairly complex workflows onto such pages, and also
assemble quite good single summary pages to report on them.
Wikis are great for knowledge sharing but they can suffer the same
feature creep as other platforms.
Feature creep isn't a problem, it's just the nature of users. You need
to manage it of course, swat the stupid requests, don't let tech
possibilities drive features, but otherwise it's evolutionary
development of user requirements. If you do it right, it's your
friend.
So when does this system stop being a wiki and start being a CMS?
Wikis aren't about pages, they're about sites. Any number of half-
baked blog engines are available that use something like wikitext to
markup an entry, but the structure of the overall site is defined by
the blog engine coding, not the content entered.
Wikis use wikitext (simplified markup, where editors may trigger
linkage) and DO NOT require additional annotation beyond this. CMS may
or may not require this, and are a superset of wiki anyway. The
distinction between wiki and non-wiki CMS is certainly grey, but
anything that has the overall site's "schema", "structure" or
"linkage" enforced from outside the wikitext alone, or requires this
to be done through anything other than an emergent behaviour based on
the wikitext, has characterised itself as being non-wiki.