Designs on the Web

New releases of Dreamweaver used to be major industry events, each version setting a new standard for web design. However, the latest Dreamweaver 8 is such a damp squib that I was worried Macromedia might have stopped development ahead of the takeover by Adobe (and hence possible replacement by GoLive). I eventually rejected this doomsday scenario, only because Adobe’s recent GoLive CS2 was even less impressive. And it’s not just this round of releases, as both Dreamweaver MX 2004 and GoLive CS were similarly disappointing.

Five years ago, web design was an exciting, constantly evolving arena, but now both main professional authoring applications seem to have ossified. Of course, both developers deny this, pointing to new capabilities like Dreamweaver 8’s XSLT-based XML-to-HTML translator and GoLive CS2’s SVG-t support for mobile handsets, but while such features may be significant in the future only a fraction of the current user-base needs them. For mainstream users producing mainstream HTML sites, where’s the new design power to give their end users a richer experience?

In one sense, a ‘rich HTML experience’ is a contradiction in terms, as Tim Berners-Lee devised HTML for marking up plain text content, not for design, but it was followed up in 1996 by a complementary design solution: Håkon Wium Lie and Bert Bos’ Cascading Style Sheets (CSS), a mark-up language for handling presentation. CSS is the key to greater web design power, as both Macromedia and Adobe clearly agree, since they’ve pushed it as the focus of their latest releases. However, I remain underwhelmed by their much-hyped ‘new’ CSS support, which turns out to be relatively minor reworkings of the functionality provided in previous releases. Building CSS support directly into the application rather than relying on hand-coding is certainly more convenient, but it doesn’t provide any more power. Dreamweaver and GoLive provide web authors with essentially the same design power today as they did seven years ago when CSS2 was released, and plenty of CSS2 features such as max-width and position:fixed properties aren’t yet fully supported.

So what sort of design power was I expecting by now? To some extent, today’s restricted design options stem from low-resolution screen displays that don’t have enough pixels to allow advanced typography, but I’d hope to be able to specify and supply downloadable web fonts for headings and display text, and also to control a host of DTP-style features from kerning through to transparency and drop shadows. The layout control offered by the first release of CSS2 is especially weak – by now I’d expect features like automatic text reflowing, multiple-column text blocks, fixed on-page frames, vertical justification, pages that rearrange themselves according to the available screen real-estate and so on. The list of features that CSS could and should tackle is near-endless.

And there are plenty of other ways the end-user experience could be enhanced. The PNG (Portable Network Graphic) bitmap format first recommended by the W3C in 1996 was specifically designed to support variable transparency for more integrated, graphically rich layouts. For vector images, the SVG (Scalable Vector Graphic) format first recommended in 2001 is especially important, potentially enabling user-interface elements like rollover buttons, graphical text (including CSS1 support) and imported graphics to be integrated directly into a page’s code. SVG also supports vector-based animation to bring pages to life, while SMIL (Synchronous Multimedia Integration Language), first recommended in 1998, enables control of audio, video and interactivity for richer online multimedia. In each case, the potential of the first specification was mouth-watering, but these recommendations are now four to nine years old – eons in web terms – and haven’t broken into the mainstream.