Suppose I want to model an email address:

  dfdl-wg@ogf.org

It’s a sequence of a string, an a character (@), and a string.

In the MBTK it might look like this:

The problem, of course, is that the parser will grab all of “dfdl-wg@ogf.org” instead of just “dfdl-wg”.  Setting validation facets doesn’t help, b/c validation happens after parsing.  Setting @ as a separator or terminator would work, but that doesn’t seem right either, as the @ is part of the email address, and I would want to keep it in the infoset.

How do I think about this problem?

TIA