xml-dev - Re: ANN: SAX 2.0 extension proposals.

[David Brownell <david-b@pacbell.net>]

David Megginson wrote:
> 
> David Brownell <david-b@pacbell.net> writes:
> 
> > However, I do wonder what you think about the fact that this will
> > need to run in environments where calls like System.getProperty()
> > aren't allowed, or which may not support dynamic loading.
> 
> Suggestions?  

Same approach used in classes like "java.lang.System" -- the API has
static class methods, but implementation of the class is unspecified.
It varies between environments; the reference implementation can be
changed.


>	I don't think that it's much use simply providing an
> interface for a factory, because then people will still have to know
> the class of the concrete implementation (and then why shouldn't they
> just know the class of the driver in the first place).

Don't be thinking that just since a factory (say) is a concrete class,
that its methods are less of an interface than a "public interface foo".

The interface/implementation line is part of the spec.  In the case
of "public interface foo" there's no flexibility:  an interface isn't
an implementation, period.  In the case of any concrete class, there's
clearly an implementation associated with the reference code -- but it's
a specification issue whether the spec mandates that implementation, or
instead says that some implementation other than the reference can be
used if it offers the same API.

- Dave

***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/threads.html
***************************************************************************