These tagclouds that TagCrowd produces are looking at word frequencies and not neccissarily what the text has been tagged with (unless you include the metadata in the text box). There are some other tagcloud makers that look at what tags have been applied.
I recently re-read this white paper on Getting Started with Controlled Vocabularies, Taxonomies and Thesauri so for fun, i through the text into TagCrowd to produce the following:
application being cases companies controlled core data developing edit existing files implementing imported information management metadata prot support system taxonomy used user values vocabulary workflow
created at TagCrowd.com