> -----Original Message-----
> From: Yves Langisch [mailto:]
> okay, to talk about semantic equivalence in general is delicate. In our
> case we have the assumption that if repeteable elements occurs, the
> order is not relevant. We just want to check the presence of data on the
> same level. At the moment we have a stylesheet which orders repeatable
> elements (can be of a complex type) by, for example, the content of the
> first sub-element. When we do this with the two input documents we can
> compare the two output documents by a string comparison (apart from
> possible whitespaces).
>
> Maybe there is a better solution?
That sounds a reasonable approach.
You probably need some sort of canonicalisation for your string comparison,
or there are plenty of "XML aware" comparison tools. DeltaXML is one, or XML
Spy has a tool built in.
Paul Spencer
