Hi Sam, This is from Michael Kay's Book, XSLT, 2nd Edition (ISBN: 1-861005-06-7), p. 802:
Xalan can use an DOM2-comformant DOM implementation as its input, and it can also attach its result tree to any DOM2-conformant document. If you supply a DOM as the source tree, Xalan will use this as it internal tree representation. If you supply input in the form of a tree, Xalan will use this as its internal tree representation. If you supply input in the form of a SAX stream or a source XML file, however, Xalan will use its own internal tree representation, called an STree. This replaces the DTM tree used in earlier versions. The STree is more efficient than a general-purpose DOM, notably because each node contains a sequence number which can be used for rapid sorting of nodes into document order. Determining document order is notoriously difficult and slow when done using standard DOM2 interfaces alone. The STree does implement all the required DOM2 interfaces, but it will bounce many of the updating methods with an exception, because it is designed to allow appending of nodes only in document order. Elizabeth Sam <[EMAIL PROTECTED]> writes: > I m trying to figure this out... > > When Xalan reads the XSL file, does it create a DOM from it > in memory ? How does the processing work ? Especially with > templates ?? > > If I understand it right, it reads the XSL with SAX and creates > a memory DOM from it (not the w3c DOM) ... ? > > Not sure what happens after that. > > How does it actually "apply" that to the XML ? > > Any pointers about the internal design would be appreciated > > Thanks! > =sam
