We have a basic User Interface (UI) issue to solve which goes a little beyond what Henry brings up. But I think the solution to the UI problem will also address Henry's issue.

In HTML, we only get one-way, one-thing links. With XML, we have a much more powerful form of hyperlink. In XML, we can create links between a word in our data (could be a document) and many other things (multi-way, multi-thing). This means we can no longer expect the simple single-click to take us where we want to go.

I suggest that we use the "mouseover" UI technique to expose links. When the user rests the mouse pointer on a word, they will see a list of links from which to choose. Perhaps the first thing they see in the list is not a link at all, but a contextual definition of the word (or phrase) if one exists.

A standard UI technique for this may already have been decided on. I'm new to XML so I don't know. But it seems obvious that something will have to be done. Everything will be "hot" in future applications*: every word, every piece of data. And not all of the links will be derived from the data's source. Our personal servers will process the datastream and add links (or take them away) before the data gets to our UI. Whether it is done by some "webby" (WBI) agent or one of Douglas' virtual terminal server processes, eventually it will be done.

*It is one of my pet peeves that users today can click on a part number in a spreadsheet and go straight into the Inventory or Purchasing application to view other data about that part. Spreadsheets generally have no clue what you are typing into them. When "everything is hot", this will be possible. I believe we may also need a completely new UI platform to support this type of environment. This gets back to my rant about computing systems supporting what people do. For instance, people name things. The system should recognize the named object in every process.

Sincerely,

Bill Bearden
BBearden@BCL.net