At 11:06 AM +0200 8/11/03, Robin Berjon wrote:
>I believe that all the position papers will be made available publicly.
That's not enough. I'm not concerned about the papers. I'm concerned
about the original data and code used to draw the conclusions in the
papers. For instance, Sun has just published an article advocating a
new format for web services based on ASN.1 that they claim is faster
than XML:
http://developer.java.sun.com/developer/technicalArticles/WebServices/fastWS/
However, I'm skeptical of a number of claims they've made. I also
suspect that their XML parsing was not nearly as fast as it could
have been. For instance, I think they used Xerces to parse documents
instead of the much faster Piccolo, thus skewing their results in
favor of the binary format. However, without access to the raw data
and code they used, I can't really tell. It's all speculation.
Without the raw data and code, this is just marketing hype, not
useful research. It can't be verified or disproven. Even if we accept
their claims at face value, all the paper shows is that Sun's
invented something faster than JAX-RPC. We won't know that they have
something faster than SOAP, unless they're willing to release a lot
more information so their claims can be tested with other languages,
APIs, and libraries.
--
Elliotte Rusty Harold
Processing XML with Java (Addison-Wesley, 2002)
http://www.cafeconleche.org/books/xmljava
http://www.amazon.com/exec/obidos/ISBN%3D0201771861/cafeaulaitA
