Consensus is the mysterious force driving open source. (The painting Consensus is by the Dutch artist Rob Voerman and was created in 1999.)
Consensus, in this sense, is a scientific concept. It's not the same as unanimity, nor is it the same as democracy. It's a common acceptance and practice which must prove its utility each day.
Standards are enforced through consensus. The W3C standards through which you view this page aren't edicts, they are common agreements implemented in your browser software.
Consensus can emerge through a political process, as with XHTML, or an economic process, as with Ethernet. Over time products which don't meet the standards are discarded. Standards can also move, as Ethernet did, from an economic to a political process, IEEE 802.3, with standards committees taking them over.
Consensus does not mean an absence of competition. There's a consensus behind SQL databases, but there are many different SQL products, and you can translate between them with ODBC. The 802.11 WiFi standards also emerged through a consensus process, and there's active competition in that market.
In open source consensus does not confer commercial advantage. Anyone can follow the consensus and derive its benefits.
A consensus supports Windows and Office as desktop standards, but because these are proprietary and Microsoft always seeks to extend them (as with its recent XML implementation) the market benefits of this consensus are subject to debate.
It is the conflict between the advantages of consensus in open source and the problem a market consensus creates (in the form of market dominance) that drives most of the friction I see on comments to this blog.
But consensus, as a force, is not going away. Its power is only going to grow. And this, as much as anything else, will continue driving open source forward.