The Future of SMIL
Webreview.com
by Kim Brown
October 8, 1999
The W3C (World Wide Web consortium) sports the following motto: "Release early, release often." It's a sage maxim to follow when changing the development course of a Web standard. The most recent release of SMIL (SMIL Boston) by the W3C Working Group demonstrates why getting a preliminary model out early is so important.
Synchronized Multimedia Integration Language (SMIL) was introduced in July of 1998. This mark-up language enables Web multimedia authors to schedule presentations so that users experience images, sounds, and text as a choreographed whole (a skill dubbers of Japanese action films never quite mastered). It also allows load balancing of data so that media can be reused. SMIL's greatest advantage, however, may be that it requires only a simple text editor to write scripts, keeping multimedia authoring as accessible as HTML.
SMIL's synchronization capability was an exciting improvement for streaming media technology. A little more than a year after its release, SMIL's impact on multimedia design is even more sweeping.
On August 3, the W3C released the first working draft of SMIL Boston. This version introduces a whole new format strategy: modularity. SMIL's functionality has been partitioned into nine new modules, all reusable extensions written in XML, each with an associated Document Object Model (DOM).
The new modules deepen the functionality within the SMIL format itself. The strongest example of this may be the Timing and Synchronization Module, with its expansion of hard and soft sync utility. Hard sync confines an entire presentation to the exact description of synchronous relationships in a time graph, while soft sync allows for a looser performance, taking into account changes in network congestion. Changes in network traffic would disrupt the SMIL 1.0 hard sync of an entire document. But SMIL Boston allows authors to apply hard and soft sync to separate elements within the document, making it much less vulnerable to fluctuations in network activity.
These new modules improve interactivity and navigation as well. Users can now navigate within the presentation, changing only part of the display. (In SMIL 1.0, either the entire display changed or a new one was created with each click.) This function is ideal for presentations with a table of contents as the tables can be integrated into the same SMIL file, simplifying authoring.
In addition to being a format for synchronized streaming media, the new SMIL modules can also be integrated into other XML-based languages such as XHTML. For instance, the new Animation Module adds animation capabilities to XHTML. Authors can create animations with JPEG and PNG images, video clips, SVG vector graphics, XHTML headlines, and other media formats. Plus, they can create these robust animations using a plain text editor. For both these reasons, the SMIL Animation Module far surpasses GIF animation, and therefore could very likely supplant it as the most popular animation type on the Web.
From an author's perspective, SMIL Boston seems headed in the right direction. However, only four of the nine new extensions are detailed in the current SMIL Boston specification, and there is much more functionality planned. The W3C SMIL Working Group may have invited feedback at such an early stage because this modular strategy has implications beyond simply authoring. Another of SMIL Boston's goals is to encourage accessibility to multimedia for the visually and auditorily impaired.
Accessibility
Since its inception, the W3C has advocated the accessibility of Web content to people with physical disabilities such as poor vision and hearing. This concern is especially pertinent when talking about multimedia content. On September 20, the W3C published the note, Accessibility Features of SMIL. As with HTML, alternative content (such as text descriptions of images) should be included in a SMIL script. Additionally, alternative or equivalent content must remain in sync with the rest of the presentation. The note explains, "Alternatives to video and audio content must be synchronized with video and audio tracks. Alternatives that are improperly synchronized may be so confusing as to be unusable."
Geoff Freed understands this point well. He is the Product Manager for the Web Access Project at WGBH's National Center for Accessible Media. (WGBH, the public television station in Boston, was the initial developer of closed captioning, which provided access to televised programming for the estimated 20 million Americans who are deaf or hard of hearing.) Freed has created several SMIL 1.0 demos for the RealSystems G2 player that illustrate how a presentation can be made accessible. His demo "car.smi" is an educational clip about the effects of Einstein's theory of relativity. The SMIL file consists of a video track, audio narration, auditory description, and text captions. The two audio tracks are timed so that the auditory descriptions do not overlap with the narration. They are also in sync with the video track, so a blind user would know exactly what content the narration is referring to. Freed's demo is also multilingual: text captions can be viewed in either English or German.
