interoperability across our work - architectural "integrity"

[chaals]

This follows comments from @igarashi50 in #120

If we believe that a spec needs to enable an interoperable ecosystem, to what extent do we believe that the overall ecosystem should demonstrate some kind of architectural integrity?

I think the point is important. This is the fundamental reason we have the TAG, as one example. At the same time, we clearly don't believe that there is "one true architecture laid down" that we have to support for everything. The discussions and conflicts between e.g. XML and JSON as data formats, RDF and flatter approaches such as XML/JSON as data model formalisms, and so on have been painful over the decades but also I think helpful.

I think we make some adjustment or offer some flexibility for what different sectors actually want to implement, and we discourage the approach that some take of trying to force a split community through diverging approaches that end up incompatible.

This doesn't seem to be clearly reflected beyond the statement "there is one world-wide web", and I think we should have more than that.

I don't have any crisp phrasing to offer though :(

[cwilso]

Related: #13