HTML 4 For Dummies - Bonus Chapters

Netscape's "dynamic HTML"


As is so often the case when giants battle for market share, the Netscape vision of dynamic HTML (notice the lowercase d on dynamic here) is quite similar to that of Microsoft. But the execution is different enough to cause headaches for Web page designers who want to build DHTML for both Internet Explorer and Navigator or Communicator users alike.

While the key concept remains one of building Web pages that can interact with users entirely on the client side, the Netscape approach is quite different from Microsoft. We’re also sorry to note that, in the eyes of most outside observers -- including the W3C -- Netscape’s vision of DHTML is less dynamic than Microsoft’s. We explain why as we go through the elements of Netscape’s vision of this technology.

In the Netscape view, dynamic HTML consists of three primary components:

  1. HTML markup, including Netscape extensions, especially those for addressing document layers through the <LAYER> tag, plus dynamic font controls.

  2. Java and Javascript, to be used in tandem with the proprietary Netscape Java-Accessible Style Sheets (JASS).

  3. An object model for HTML documents, which differs substantially from the Microsoft DOM.

The most obvious distinction between Netscape and Microsoft weighs heavily in Microsoft’s favor -- that is, Microsoft incorporated standard notions of style sheets (CSS) and standard HTML 4.0 markup into its version of DHTML. Microsoft has also offered its object model as the basis for standards development to the W3C, where it’s been taken up with considerable enthusiasm (but not outright adoption). Netscape, on the other hand, has built its vision of DHTML around proprietary markup, and a completely proprietary, nonstandard version of style sheets, plus a rather more limited notion of acceptable scripting languages (JavaScript only).

Netscape's dynamic HTML comes in <LAYER>s

The Netscape <LAYER> tag anchors its implementation of DHTML: It provides the mechanism used to position HTML content into layers. It also provides the methods used to impose ordering on these layers, so that their precedence, visibility, and display characteristics can be manipulated in the browser’s display. A similar x, y, z coordinate model is used to reference HTML elements and objects and to control their precise positioning on a Web page that not only refers to the elements' position horizontally and vertically on the page, but their positions in front and behind each other in layers of data.

The layering model is common with many high-end graphics programs, such as Adobe Photoshop. A discussion about layering can take up a whole chapter of a book and is too complex for this book. Suffice it to say that, with layering, the only elements that are displayed at any given moment are those whose visibility attributes are set to require their display, and whose position places them within the frame of the current display area for the page, whatever that might be.

As HTML markup goes, <LAYER> is a powerful and highly descriptive tag, replete with numerous attributes and associated values. Some of its most noteworthy attributes are as follows:

Taken as a whole, the layer mechanism gives page designers a great way to break content up into mutually exclusive sets of data, text, and graphics, and to switch among sets as needed. This functionality provides the foundation for the rest of Netscape’s approach to DHTML.

Java-Accessible Style Sheets (JASS)

Java-Accessible Style Sheets (JASS) is intended to let page designers access any object with a NAME attribute while an HTML document is loading on the client side. JASS works only with JavaScript at present and with programs that can generate JavaScript output.

The reason why we say that the Netscape vision of DHTML is less dynamic than that of Microsoft hinges on one limitation of JASS: It does not permit dynamic changes to page content or structure after the page has finished loading. (The Microsoft implementation enables you to operate on any object element in the all object at any time while the page is loaded into the browser). This limitation means that all views of data must be built in advance and the layer mechanism must be used to shift among a collection of alternate, but static, views to create dynamic behavior.

The Microsoft approach is more direct and allows the data to be changed and redisplayed directly at any time. The Netscape approach does not entirely bar a program’s manipulation of page, but it does prevent programs from acting on Web pages after they fully load. Any programmatic behavior to be imposed on Web documents must therefore be handled before the page is downloaded, rather than applied as needed on the client side. While this requires more forethought, it also ensures faster performance (because switching from one set of layers to another is faster than reflowing text on screen).

The real downside for Web page designers is that what constitutes DHTML markup for IE users is different from what constitutes DHTML markup for Netscape users. Ultimately, designers must create two sets of markup to deliver the same dynamic effects to users with IE and users with Netscape.

Shelley Powers wrote a great article for NetscapeWorld entitled “Writing Dynamic Pages That Look and Work Great in Both Navigator and IE.” This piece does a good job of describing how designers can build pages that can use DHTML but work equally well for both kinds of users. You can find this article at

Database experts like to talk about something called the duplicate data problem. This term describes what happens when two copies of the same data set must be kept synchronized -- a process that involves many headaches. Unfortunately, the same aching sensations may accompany your effort to create and maintain duplicate sets of markup for IE and Navigator users that deliver similar dynamic HTML behaviors.

Netscape object model

The Netscape object model uses a similar notion of organizing and addressing elements within an HTML document. But the notation is completely different from that of Microsoft and depends entirely on unique names being assigned to individual elements to make the elements addressable. Thus, object references look much different, and not everything is guaranteed to be accessible (unless it has a name). For those reasons, Netscape’s object model has not been widely adopted outside the Netscape developer community.

Rounding out the Netscape vision of DHTML

Though the Netscape vision of DHTML differs from, and is arguably less dynamic than, the vision of Microsoft, Netscape’s approach still offers similar capabilities and functionality. Seeing the effects that this technology can produce provides the most compelling demonstration of its features and benefits. Table E3-2 includes further information resources on the Netscape view of DHTML, plus some great examples of this technology at work.

Table E3-2: Netscape DHTML Resources and Examples

Category Description URL
Overview Dynamic HTML Overview
  Compelling Content
  Open Standards Guarantee
Ads Online fashion mall
  Premiere Tech’s Layers demo
Animations Taboca ArtworK
  Sinfomic Layers Demo
  Games Mix-a-Pol
  Ridge Layer
  Space Fire
  Canvas Mode Stella Chelsea Demo
  Style Sheets Royal Newsletter Demo
  Web Design Group Tutorial

Extra 3 Main Page | Previous Section | Next Section

Home | Bonus Chapters | FTP Resources | Site Overview | Book Contents | Book URLs | Book Examples | Wayfinding Toolkit | Contact Info

E-mail: HTML For Dummies

Webmaster: Natanya Pitts, LANWrights
Copyright Information
Revised -- January 16, 1998