One reason alone sinks the idea of colored text, mentioned in Mr. Cross' own paper: it discriminates against a minority. Besides, colored text cannot be controlled, since colors displayed are ultimately user configurable. There are even cultural biases against writing in certain colors! Ultimately one would end up being a standards body deciding what color schemes would be acceptable to be certified.
A whiter shade of gray?
Subtle shades of highlighting behind text would probably work better, but suffers from some of the same drawbacks. Punctuation would soon be overused and distracting. Ms. Werner mentions "see behind," so another alternative may be that text or highlighting can be turned on or off behind the current text.
Leaving The (confidence) View?
Mr. Cross tested a system whereby the user must consciously select the "confidence view," but I would think there would be: a) rampant non-use, and b) the simple fact that sites such as Wikipediaetal can never have High Confidence, nor should they. If I were a teacher and a student submitted a report based on articles from Encyclopedia Titannica, rather than agreed-upon authoritative sources, they would be engaged in a titanic struggle to raise his/her grade. The purpose and reason for the success of this paradigm has little to do with knowledge, and a lot to do with getting noticed.
Old but still true?
The author gives some general criteria for determining confidence levels, distilled down to their essence, consisting of peer review, and since "well-documented" is contrasted with "brand new," I would consider another criteria to be age, implied also in Mr. Cross' paper, and whether something has sufficient documentation to be considered pretty true. The last criteria carries the most weight in my opinion.
Cyclical time is merely human, the arrow of time is cosmic...?
One objection is that there are such things as very weighty opinions, and even if Steven Hawkings voluntarily submitted a paper for certification, he might think that more of his "opinions" should be marked "High Confidence" out the gate. Also, in order for the Ribbon to mean anything, it would have to be a widely used and sought-after standard. The majority of content-creators are more interested in profit than reliability, and there are other social factors as well, when you think of why someone would create a wiki or a blog, etc. Confidence should probably not be applied to the Web in general. Although I may want one Web site of mine certified, I would not want my blog rated at all. Perhaps no one would read it or comment on it, like they do now. I hope I am not misrepresenting or misunderstanding the intention.
Do search engines dream of artificial sheep?
Finally, there should be a way to quantify "confidence" mathematically so that confidence can be automatically generated. Most people now use search engines as portals, so these engines should have a mathematical algorithm to give a confidence score rather than the "page rank" popularity contest. Here we are on the border of AI, I think, so although confidence is a worthy goal, my opinion is that basic changes are needed to the very structure of communication, computing and how people, in fact, are, to make this a reality.