There’s a super-entertaining comment thread/flamewar going down at Ajaxian in response to Dion Almaer’s recent editorial on The Future of CSS and the end of 3.0. Almaer has lots of interesting and perceptive things to say about the maddeningly erratic way in which new standards make their way into general use. But the vast majority of commentators jumped at the chance to sound off on some of his minor points:

CSS is great for simple web style. CSS is awful for layout. Rich Ajax apps need layout. You spend the majority of your time trying to get CSS working correctly!

Of those four assertations, there’s not a single one I can fully endorse. IMHO, current implementations of CSS are _pretty good_ for simple web style, but they’re not great. (See the comments for many examples: rounded corners, varied fonts, etc.) They’re not great for layout, either, but if you take the time to learn the tricks – and account for spotty browser implementations – then you can develop attractive, robust and usable UIs using CSS. As for whether rich Ajax apps need layout, well, yes. But what kind of layout is open to considerable interpretation. Just because the first few generations of web designers kept trying to apply the same old print paradigms to the browser doesn’t mean it was right. The same goes for the current wave Ajax frameworks and the desktop-app layout paradigms some of them attempt to implement.

The assertation that bothers me the most, though, is the one about the time it takes to get CSS working properly. As with any code executed in the browser, CSS is subject to the quirks and peccidillos of browser vendors, operating systems and ever-mutating standards. Still, once you account for the major differences in execution environment, it’s not that hard to come up with a single global stylesheet that can zero out built-in styles and browser quirks to offers a more or less blank canvas for cross-browser CSS development. CSS authors may not release their design frameworks under open-source licences the way JavaScript authors do. But that doesn’t mean they don’t exist. Just peek under the hood at a big, forward-thinking consumer webapp or pick up a Jeffrey Zeldman book.

Regardless of what _I_ think, the thread at Ajaxian provides an interesting vox populi. As can be expected, comments range from wholehearted endorsements of Almaer’s position to finger-wagging about the importance of web standards to chest-puffing from people who’ve mastered the tricky art of CSS layout and don’t have much sympathy for those who haven’t. I’m fascinated by the way participants in these conversations let their personal investment in a certain skill set lead them to make sweeping generalizations about entire swaths of the software-engineering world. Just because you know how to make a site look a certain way using tables, or floats, or XSLT, doesn’t mean that that’s the only way, or that the Internet won’t eventually cough up a much better way than any of the above.

The fact is, any programming language or development framework can be improved, and there’s lots to both love and hate about the CSS specs and the various imperfect implementations of them. It’s _so_ far from the all-or-nothing proposition that many of the commentators would have you believe.

Dealing with client-side technology is frustrating for anybody who’s used to having complete control over their execution environment, or a stable platform at all. The reason the pace of change is so glacial on the client side is that standards take time to develop; vendors take time to implement them; and a variety of market forces contribute to how well and how quickly they’re implemented. Nobody said the Open Web was easy, but it’s vastly preferable a host of closed-off, proprietary formats controlled by individual vendors and immune to grassroots innovation.

Any developer who’s lived through successive generations of client-side technology knows that things have only gotten better with time. It’s hard to imagine that standards won’t continue to improve, no matter how painfully slow the process.

Share/Bookmark