Lists Home |
Date Index |
Gavin Thomas Nicol wrote:
> > 2) After we had namespaces, and you chose to use them, you made an
> > assumption about the data type of the element based on its name, which
> > was "guaranteed" to be universal.
> The names cannot be "guaranteed" to be unique... especially over time.
It seems there are two guarantees here. The first is that somebody can
easily steal your names. The second is that the definition (semantic and
structural) of an element type can be changed at random and that the
longer an element type exists, the more likely this is to happen. I
assume that you're referring to the second? If so, schemas will help
with structural changes (and an improvement over the non-schema
situation) but nothing is likely to help with semantic changes.