Re: Client-side highlighting; tag proposal

Steven J. DeRose (
Mon, 13 Mar 1995 15:00:56 +0500

At 1:19 PM 3/13/95 +0500, Steinar Bang wrote:
>Or, not to be limited to element boundaries: you could identify the
>container element this way, and then just give the byte offset for
>start and stop in the content of the element.

How about relating this to a solution for tree addressing mentioned
in comments in the DTD? As long as you can specify a walk down the
tree to the element containing the start (and likewise the end), then
just give offsets within it, it's pretty reliable (it's mildly interesting
to characterize the precise classes of edits that do and do not invalidate
such a locator, but I'll leave the exercise to the reader...).