[
Lists Home |
Date Index |
Thread Index
]
Manos Batsis scripsit:
> However, this would seem a lot more logical if some other form of syntax =
> was used; it is redundant because it's XML in the first place.
Correct. The main advantage of the XML syntax is that it already has a
mechanism for binding QNames for elements and properties: with an ad hoc
syntax, I'd have to reinvent that.
> My question is, you said this is a serialization for all infosets, not =
> just the infoset. So, suppose I begin to erase the explicit =
> serialization towards the original document, where should I stop to have =
> enough information for any infoset and not just the infoset? In other =
> words, what is that piece in the design that makes this serialization =
> suitable for any infoset?
Because infosets, at least the ones we have seen so far, consist of
information items which have properties, and a property is either a
simple type (boolean, integer, string, enum) or it's an unordered set
of information items, or an ordered list of information items. In
short, it is a graph, and Layman normal form is one way to represent
graphs as XML.
--
John Cowan <jcowan@reutershealth.com>
http://www.reutershealth.com http://www.ccil.org/~cowan
Yakka foob mog. Grug pubbawup zink wattoom gazork. Chumble spuzz.
-- Calvin, giving Newton's First Law "in his own words"
|
