[XML-SIG] getDOMImplementation() extension

Fred L. Drake, Jr. fdrake@acm.org
Tue, 5 Nov 2002 14:37:56 -0500

Martin v. Loewis writes:
 > I like it (obviously): errors should never pass silently, unless
 > expliclitly silenced.

I agree.  The real difference between the W3C idea of
getDOMImplementation() and the one we're using is in whether not
getting an implementation for some particular feature set is an
error.  See Ray Whitmer's response to my questions about this on the
www-dom list:


Which is the "right" thing is rightly very application-specific.  For
the W3C flavor of this, there's no such thing as "just gimme a DOM";
you always ask for some specific set of features, and the application
can reasonably say "ok, that didn't work, let's try a reduced feature
set and offer reduced functionality" (where any decent application
will be able to tell the user if it makes sense to do so).

As I consider in my response to Ray's comments:


the right approach may be to have a second function (either somewhere
else or with a different name) that does implement the W3C API for
this, but the overloading of the name may well not be the right way to
go about this.


Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Zope Corporation