By pelegri on Jun 11, 2006
It is hard to correct misconceptions, but let me try:
Fast Infoset does not require a ASN.1 toolkit.
The misconception probably arises because the ITU-T Rec. X.891 | ISO/IEC 24824-1 Specification uses the ASN.1 descriptive formalism and because it was done by the same ISO / ITU-T group that worked over the years on ASN.1, but there is no required dependency - for instance the Implementation in Project GlassFish does not use ASN.1 at all.
I'm mentioning this because of Paul's latest blog on Fast Infoset and Scientific Computation where he reviews an article on Binary XML for Scientific Computations which, unfortunately get this part wrong. Please help us correct this misconception.
There has been good recent progress on Fast Infoset, including More Implementations. Paul also wrote recently about API Improvements for External Vocabularies and Encoding and Decoding Primitive Types.
The picture is one of the Cray 2 of of the classical-style Super Computer used in complex scientific computations. A different tackle on handling computationally intensive problems focuses not just on the actual run-time but on the design-to-result time in DARPA's High Productivity Computing Intiative, including Sun's Efforts.