Thanks for the effort to keep the taxonomy updated! It must be a chore to orchestrate 1000 opinions!
One of my concerns with a taxonomy is not to create logical do-loops, which have been the downfall of more than one architectural reference model in the past.
I would propose 2 changes, one of which is simple.
The simple one is to replace the term "data center" with "computing facilities". While the term may be correct as defined, it is widely used by non-TBMers to mean "all the capabilities that such a facility typically houses." In the organizations I have ben working with, this is the number one source of confusion and holding to it as defined in TBM frankly threatens the credibility of the framework in the eyes of the people who are really doing the IT work.
A more complex one may or may not be a proposed change, depending on how the language ends up, so I'll state it here in terms of what I think we need to shoot for.
With regard to business services, shared services and platform services. Again it is matter of making the model clear enough that regular IT managers can grasp it, as well as maintaining comparability across organizations and even sub-organizations. The IT Services basically represent the manageable layers of the OS stack and conform to the majority of EA structures that already exist. Until you get to the business services at the top of the model, everything else below it is function-agnostic. We should not start setting up separate service groups for things that really belong in one of the other service groups; that is where the death spirals begin.
.
My suggestion would be to retain business applications services as just what it says: applications that deliver specific business functionalities. Now I know that we are not casual observers, but most IT and finance managers are (at least with regard to TBM) and they are the ones who will make it work or fall apart as they sustain what we put in place. When the casual observer sees "business services" without the "applications" word they can get confused about where we are in the architecture, and they tend to think of an end-to-end business process [purchasing] vs [purchasing management software]. If they are thinking IT at that point, they are thinking "system" because that is what they have contact with: a purchasing system that includes end-to-end IT (application, platform, infrastructure and all) and the next thing you know all the costs of those other services will be lumped in there, putting us right back to where we were before TBM tried to break them out. This what led to the practical demise of the last round of the DoD reference architectures: including entire systems at the "business applications" level.
Adding the Shared Services could work if we wanted to distinguish IT functional applications (such as dev-test or collaboration tools) that a person could use in support of any business or IT function (i.e. Shared Service) from Platform IT, meaning afunctional solutions that users can't "use" directly (such as enterprise service bus, hosting layers, database clusters, middleware etc for any business or IT from user-invisible applications that manage infrastructure.
I guess that is a lot more than 2 cents ;-)