"dot oh" = version controlled terminology used as marketing speak to describe a concept
The concept may be a new one or it may simply be new to the marketing department. Most importantly, advocates want readers to believe it is new. The underlying concepts can be very meaningful, especially when one tries to compare trends (as in SOA versus Web 2.0). I just cringe at the shrink wrapped packaging of fuzzily defined terms. So let's talk about the meat on the bones. Here is the run down.
SOA 1.0 - client/server service oriented architecture. This is what we know today as SOA including soap/web services.
SOA 2.0 - event driven service oriented architecture. Defined as "a software architecture that defines how systems can be engineered and designed to sense and respond to events."
Web 2.0 - Collaborative, fast development of applications. RSS/Atom, mashups, wikis and the like. Often uses existing technology in a new way to add value.
Web 3.0 - Semantic web. RDF and related semantically based applications.
I think these terms really have meaning when it comes to comparing trends. Comparing SOA (1.0) versus Web 2.0 is an interesting one. The idea behind these indicates two different views of how applications are to be developed. Take data integration (something I know a bit about). The traditional route of thinking is to create a data model of a business document, express it in Xml Schema, and ship it to my trading partner via services as SOAP (or message que or ...). This reflects a certain architecture and modeling that is not trivial. In a web 2.0 world, you could have a data integration in simpler terms. Use Xml and share it with trading partners via simple https or even syndicate it via RSS. The former is basically architected and the latter is mashed together. This reflects an interesting debate about how to develop apps and integrate systems. Each has their own strengths and I think will both have their niches, so I think it is a false choice.
Comparing Web 2.0 with Web 3.0 is also an interesting one. The web2.0 is intended to be quick, easy, leveraging existing apps, and collaborative in nature. The idea of a semantically based web (3.0 here) is basically the opposite of this. Many will disagree with me here, but creating a semantic web is very difficult and involved alot of analysis which is the antithesis of the quick mashup. Entire companies do nothing but mine business data in a semantic web based fashion. So here again each of these concept has their own niche.
As a side bar, I am an old librarian/archivist and so am sympathetic to semantic web technologies. I looked at RDF a long time ago and have always wanted to justify using it in my work, but never had the opportunity. I even had a thought to use it at HR-XML. But I always kept getting caught between what was the ideal thing to do and what was the reality of current demands.