@LKBrwn_DBA: I hope that the article you linked to was meant only in jest. If it is sincerely intended then IMHO it appears written with a myopic perspective. [Or maybe it is dated - many arguments it makes do not hold in the present times, either for Oracle or for XML.]

Just so that it does not misguide people who miss the joke and are seriously considering using XML DB in Oracle, here are some counter points.

By jamming all of those tags into the data you can easily more than double the size of the data files being sent across the network. And the bigger the file, the slower the transmission time.
Whenever data is transferred between disparate systems, it must be structured in some way that is understandable to the communicating systems. Structured information contains both content and some indication of what role that content plays. Tags provide structure and meaning to data.

The article does not propose an alternative approach to data transfer that provides the key advantages of XML - flexibility, simplicity, interoperability - without the use of tags. It is the problems faced by older complex data transfer mechanisms like EDI that led to the evolution and increased acceptance of XML.

Bottomline: The benefits of using XML in the right context outweigh the extra storage space it takes up in the database.

The first approach is to be preferred. It is less troublesome to administer but more complicated to implement. It requires someone (the DBA?) to map each tagged element to an appropriate column that already exists in a table.
For a developer (the DBA?) who wishes to remain in the comfort zone of familiarity, the first approach would indeed be preferable. A developer (the DBA?) who learns that this can be done fairly easily in Oracle would make a more intelligent, rational choice based on what suits the application best, rather than what entails least work for him.

And, once again, that data is of no use to anyone who does not have the parsing program. So the database is holding data that is not really defined in any way other than as a giant clump of text. Hello? Does anyone really view this as progress?
Oracle has the capability for easy traversal, access, manipulation of the "giant clump of text" that is XML using SQL/XML or XML-specific packages - the parsing program is right there - and we do view that as progress.

XML offers little control over data type and length
On the contrary, current versions of XML Schema languages allow sophisticated constraint definitions, including datatype specification (both primitive and user-defined), length, minimum/maximum occurences, default values and so on.

XML stands for eXtensible Markup Language. And that is another thing wrong with it – I hate acronyms that use letters from the middle of words!
The acronym is derived phonetically. The primary letter from the first syllable of each word is chosen: the X sound of 'eXtensible' is arguably more pronounced than the E - which intuitively sounds good to me. At least I prefer XML to EML
[How would the world be if we chose technologies based on our biases for their names? I would stay miles away from C# because I hate names with special characters ]