Lists Home |
Date Index |
Many thanks to Joe, Len, and Eric for their comments. I would like to
follow up with a comment/question on complexity in building information
First, some questions about Len's comments:
Len Bullard wrote:
> Don't forget attributes.
Is an attribute a "building block"? Or, is an attribute "in-lined metadata"
on a tag - analogous to scribbling a note on a Lego building block to
describe its content?
> Also, in practice, building up complex systems requires namespaces
Is a namespace a "building block"? Or, is a namespace simply more metadata
on a tag - analogous to coloring a Lego building block to distinguish it
Let's suppose that we consider attributes and namespaces not to be building
blocks. Then XML is able to accomplish all of its complexity with just 2
building blocks and 2 assembly mechanisms. Does this not tell us something
about the maximum number of building blocks and assembly mechanisms that an
information system should have? Does it not place an upper bound on the
complexity required of an information system?
With XML, information structures of virtually infinite variety and
complexity can be created. And XML is usable across a broad range of
applications. Surely, then, an information system - which undoubtedly has a
much narrower focus than XML - should be able to achieve all the variety and
complexity of information structures that it needs with no more than 2
building blocks and 2 assembly mechanisms. Thoughts? /Roger