Joseph Kesselman wrote:
It's true that object serialization/deserialization is not a good choice for XML tree structures - in my tests Java serialization/deserialization for document models takes 2-3X the time of serialization to/parsing from text (and the data size is also much larger).The internal structure would be something like a DOM tree, which tends to take more bytes than the XML stream does... and object serialization/deserialization is often not a heck of a lot faster than XML parsing. Most of the proposals for "binary XML" representations have hit that same set of rocks and been abandoned.
There definitely are more efficient formats for exchanging document Infosets than text, though. See my XBIS (http://www.xbis.org) for one example. XBIS uses a compact format for representing the complete Infoset of documents. I was able to get 6-8X Xerces parser performance for input (and roughly equivalent gains on output) over XML text using the SAX2 adapter. I'll actually be updating the XBIS code with adapters to/from DOM/JDOM/dom4j soon, and with comparisons to zip performance.
The only rocks I've run into with this are the ones that say it's okay to gzip XML documents for transmission, but any other form of transformation is verboten... :-)
- Dennis
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
