- From: Jimmy Cerra <jimbofc@yahoo.com>
- Date: Sat, 27 Apr 2002 22:20:05 -0400
- To: <www-dom@w3.org>
I was reading up on the current state of XPointer, XPath and etcetera. It
occurs to me that, if XPointer ever becomes a Recommendation (1), then it can be
used to generalize getElementByID() into a function which can return any point
in a document. Say we have a function called getElementByXPointer(). Since
XPointer supports ids (2), then:
] getElementByID('1701A')
...could be represented using XPointer as either:
] getElementByXPointer('1701A')
] getElementByXPointer('/1/1/2')
] getElementByXPointer('fed/2')
...in the following XML (3):
] <?xml version="1.0"?>
] <?xml-id names="id xcha je"?>
] <starships>
] <alien id="fed" xml:lang="en-US">
] <ship id="1701" />
] <ship id="1701A" />
] <ship id="2000" />
] </alien>
] <alien xcha="klingon" xml:lang="x-klingon">
] <ship xcha="IKS7" />
] <ship xcha="IKS38k" />
] </alien>
] </starships>
This seems to be a natural generalization of the getElementByID function; it
provides more flexiblity and brings some additional functionality to the DOM (an
XPointer can return a single element or a set of elements/nodes). Is this being
considered for the next version of the DOM?
---
Jimmy Cerra
---
1) I really hope XPointer becomes a recommendation, as I like some of the
features, such as Child Sequences.
2) AKA Bare Names.
3) Note that I used a processing instruction to tell you what an id is.
Received on Saturday, 27 April 2002 22:20:12 UTC