Lists Home |
Date Index |
Dare Obasanjo wrote:
> Exactly how is the XML infoset supposed to be represented and/or accessed by
> application programmers?
It's deliberately left unspecified in the Infoset spec.
The intent is that specific APIs document what parts of the
Infoset they support and how it's supported.
> Since the infoset seems to mirror the DOM to some
> degree will the primary usage model be some sort of DOM-like API with current
> DOM implementations being given the ability to support it with certain
There is no more a primary usage model for the Infoset
than there is for XML itself; DOM, SAX, and XSLT all
support the Infoset, but in very different ways.
> Or is the XML Infoset recommendation supposed to exist merely as
> reference for use in discussions between experts and others with domain
> knowledge without it actually having any implementation impact on users of XM
> in general?
More or less. The Infoset REC is mostly for design
and documentation purposes. It provides a common vocabulary
so programmers can tell (for example) that what the SAX
documentation calls an "element information item" is the
same thing as what the DOM documentation calls an "element
> Finally, the XML infoset is particularly DTD centric with little if any
> mention of XML schemas but on the other hand I keep hearing about the PSVI
> with regards to schema awareness and data types, is there any reason for this
> disjointedness or is there some discussion or document I can be pointed to
> that clarifies the issue?
REC-xml-infoset includes a [document type declaration] information
item because <!DOCTYPE ...> declarations are part of XML 1.0.
REC-xmlschema-* defines new information items that schema-aware
processors must also support (i.e., a PSVI). These new information
items aren't part of REC-xml-infoset because they aren't part of
XML 1.0. Does that clarify things?