The Action Sheets concept, as published in the W3C Note, is outlined in a very vague way, which has the advantage of allowing a lot of room to maneuver - or, if you prefer, the freedom to make up damned near anything you want. This isn't to say that there's no structure. The heart of the Action Sheets proposal lays down the idea that any Action System will require four components:
- Selector syntax - those parts of a document which can be
specified by a selector; for example, CSS1 selectors
- Selection domain - a way of declaring which parts of the
document will be affected by the various scripts; for example, the set of
HTML elements or a set of XML elements
- Action domain - the allowed actions, or events, which can be
used to trigger scripted behaviors; for example, onMouseClick
- Action definition - the scripts themselves
The Note then goes on to say that these areas are "fairly independent from
one another," which seems to be another way of saying: "Consistency between
these areas is not required." Of course, given the extreme amount of
freedom this system allows, an Action System with internal consistency
could be quite powerful.
Somewhat remarkable for a proposal from Netscape is that Action Sheets are
not tied to JavaScript in any way. It is certainly possible to
reproduce the entirety of JavaScript's capabilities within this system, and
even enhance them, but one could also define Action Systems for the
inclusion of VBScript, Perl, awk, Lisp, or just about anything else.
Another interesting aspect of Action Sheets is that they can be created by using eXtensible Markup Language (XML), therefore making them easy to understand and share. Thanks to their use of XML, an up-and-coming W3C standard, Netscape has allowed for unlimited future expansion as well as a format which other implementors can easily follow. Microsoft's DHTML Behaviors files also use XML documents, and so share the same benefits.
next page»
