Availability of Fast Infoset

["Alessandro Triglia" <sandro@mclink.it>]

Hi

I am pleased to post the following information about the status of
availability of Fast Infoset.

http://fi-interop.dev.java.net/

"Fast Infoset is now available and interoperable on a wide variety of
platforms such as Microsoft .NET and .NET CF, Sun GlassFish, BEA WebLogic,
Java SE 6, IBM SDK for Java v6 Early Release, TMax Soft JEUS 6, as well as
Linux, Solaris, and Win32."

Fast Infoset is an International Standard jointly published by ISO/IEC and
the ITU-T, which specifies a serialization of the XML infoset, and can be
used in certain applications as a fast and compact alternative to XML 1.0.
Fast Infoset preserves all the properties of an XML document that are
reflected in the XML infoset.  

Note that Fast Infoset only supports XML documents that are well-formed and
conform to the W3C Namespaces Recommendation.  In fact, it supports all such
documents with the exception of certain "extreme" infosets (e.g., those
containing single attribute values longer than 4 gigabytes, or single
comments or processing instructions whose content is longer than 4
gigabytes).

Fast Infoset does not support documents that are not namespace-well-formed.
Also, even though it supports synthetic infosets, it only supports those
that could have originated from XML documents (that is, any infosets that
are impossible to serialize in XML are not supported).  Also note that,
because Fast Infoset is a representation of the XML infoset (as opposed to a
representation of XML 1.0), it does not provide a representation for certain
features of XML 1.0 (e.g., the declaration and use of entities) that are not
represented in the XML infoset.

Aside from the restriction relative to "extreme" infosets mentioned above,
Fast Infoset is usable with XML Schema validation, XPath 2.0, SOAP 1.2, and
other technologies that are defined in terms of the XML Infoset (or
something derived from the XML Infoset), rather than in terms of XML 1.0.

(For example, in order to schema-validate a Fast Infoset document, one
must--conceptually--first convert the Fast Infoset document to an XML
infoset and then validate the infoset.  This is analogous to performing
schema-validation of an XML 1.0 document, which must also
be--conceptually--converted to an XML infoset, because the XML Schema
Recommendation specifies validation in terms of an XML infoset, not in terms
of an XML document.)

The above implies that it is possible to use Fast Infoset in many situations
in which XML 1.0 is used today, and the impact on the other components of
the system should be minimal.

(For further clarification:  Fast Infoset is not a lossless representation
of XML 1.0--it is a lossless representation of XML infosets;  this implies
that any technology defined in terms of the XML infoset can be used equally
well with XML 1.0 documents as with Fast Infoset documents.)

The Fast Infoset Recommendation can be downloaded for free from the ITU-T
website at:  

http://www.itu.int/rec/T-REC-X.891/en

Alessandro Triglia
OSS Nokalva