Lists Home |
Date Index |
On Friday, February 7, 2003, at 11:06 PM, Bawcom, Aaron wrote:
> Once the user has specified an XML Schema, there would be a root node
> to start with.
> When the user right clicked on the root node, the user is presented
> with a context menu of *only valid elements and attributes" that could
> be specified at that point. So basically, the possible elements and
> attributes you could add as children or properties of the current node
> would be dynamically determined by the Schema that was loaded.
Mark Sifer, then of IBM Haifa, made a research editor called "Xeena
for Schema" For some reason IBM withdrew it quickly, but some
quick people still have copies :-)
It went further than you say: rather than clicking on the root showing you elements at
a single level, it would let you navigate multiple sublevels. Say your top element
was html you could drill down to html>body>div>h1, say: the editor would then generate
all the intermediate (and required, I think) elements to that point. This is a fast way
to create elements fast.
The bane of XML editing is that you so often have to make repeated choices, labouriously
being presented with choices. That is OK if the DTD is fairly flat, but if the schema is
rather deep it drives you mad.
Dr Sifer also had some very interesting other ideas in Xeena for Schema: in particular,
a really neat visualization way for viewing data by multiple categories.
By the way, I think as a generalization, "In a perfect world, the XML Editor would be
a tree view which would represent the actual XML document itself" is dead wrong,
and looks like a marketing pre-announcement :-) For example, if your document is not
yet WF or grossly invalid, trees get in the way.