Musings about Adobe Muse

Do you remember Dreamweaver? Of course you do. Some memories refuse to die quietly.

A decade ago, Macromedia’s Dreamweaver was one of the preferred weapons for building websites, having inherited the dubious throne from the infamous Microsoft FrontPage. It allowed ordinary mortals to create something that, when viewed from a safe distance and in the correct browser, vaguely resembled a website.

Then you looked at the generated code. It wasn’t pretty.

Dreamweaver was eventually acquired by Adobe, who for reasons unknown continued developing it. And now Adobe is back with another attempt to liberate us from the apparently unbearable burden of understanding the medium we’re working with.

Adobe Muse is based on Adobe Air and lets you drag and drop components onto a layout pane much like InDesign. No coding required. Just arrange the boxes until everything looks nice and let the machinery underneath worry about the consequences. Which may sound wonderful. So does an all-you-can-eat seafood buffet at a motorway service station.

The problem is that Muse appears obsessed with fixed dimensions at precisely the moment the rest of the web development world is talking about Responsive Web Design. The web is becoming fluid, adaptable and increasingly indifferent to the size of the screen displaying it. Meanwhile, Muse seems determined to nail everything firmly to the table.

Despite the name, I’m not convinced Muse will inspire anyone to greatness. Divine inspiration may arrive in many forms. Bernini had visions of angels. Michelangelo had the Sistine Chapel. Web developers get 223 div elements.

Divine inspiration may come in many shapes, but always inspect the source.

Adobe claims the generated code is standards-based. Naturally, I had to look. I inspected a few Muse-generated sites, including Adobe’s own Muse site, and suddenly found myself transported ten years backwards in time, staring once again into the dark, steaming entrails of Dreamweaver-generated markup. The horror, as Kurtz would say.

Progress is a funny thing. Half a decade ago, people were finally crawling from the wreckage of the 1990s and abandoning table-based layouts. We congratulated ourselves on becoming enlightened and semantic, then proceeded to replace “tag soup” with “div soup”.

It is now 2011, and Adobe’s relatively simple Muse website contains 223 div elements. Two hundred and twenty-three. Among the delicacies being served is this: (sensitive readers be warned)

There are crimes committed in dark alleys with more dignity. And this isn’t merely about developers being precious about beautifully formatted source code. Good markup isn’t some private fetish practiced by standards geeks after midnight. It affects browser compatibility, loading times, accessibility, search engines and maintainability. All that boring plumbing that determines whether an attractive website actually works for the people unfortunate enough to visit it.

I have no objection to tools that allow people to create websites without becoming professional developers. Quite the opposite. If someone builds a tool that lets anyone create something beautiful, accessible and technically sound without having to know the difference between an element and an attribute, that’s wonderful.

But hiding complexity is not the same thing as solving it. A machine that produces bad code more conveniently is still producing bad code. It just lets you produce considerably more of it before lunch.

Maybe Muse will improve. Maybe one day visual web-design tools really will generate clean, semantic, accessible markup while allowing designers to concentrate entirely on their ideas. That would genuinely be something worth celebrating. Until then, inspect the source before ordering dessert.