Twitter, ye markup be non-standarrrrrd.
Filed in: Web, design, vent, Fri, Sep 19 2008 11:10 PT
Twitter unveiled a new redesign today. Very little of it is really noticeably different, until you look at the underlying code. Which, last I recall, used to be pretty good–they even used the fancy-pants XHTML 1.0 Strict doctype (though still using tables for layout, which the spec says it shouldn’t do).
But one thing about the latest version makes me wonder just what the hell they’re doing these days.
<center>.
The <center> element. In September 2008. From the “it” Web 2.0 company. Seriously.
I know this will make me sound like the annoying standardista, but how could anyone who still uses <center> still be doing web design professionally in, of all places, San Francisco? This is an element which has been deprecated for eleven years. Do we really have people who haven’t changed their coding practices since before 1997?
Sadly, yes. And the worse news is, they’re writing books. I just saw a book whose first edition was published in July 2008, which teaches users to use <center>, to do layout tables, to use CSS primarily just for font selections, and loads of other outdated guidance. This is material from the bad old days of web design, and it simply gets regurgitated over and over again. To quote the late, great George Carlin, “it’s all bullshit, and it’s bad for ya.”
I don’t know what it is going to take to finally cull the proverbial herd of these kinds of authors and designers. But each time I see this, it makes me wonder when we can expect some kind of professionalism out of the average content producer. Many of us have been talking about this stuff for years. It’s de rigueur at many web conferences, to the point that people now roll their eyes at it. And yet, it continues. I also don’t know whether Twitter is doing this in-house or if they hired an external designer. But certainly, somebody there dropped the ball.
And I know that one <center> is not a big thing. It’s just a symptom of a larger disease: that of lazy, ignorant and/or incurious designers. When someone sticks to one way to do something without ever updating their own skill set, their designs get more and more inflexible. Which makes redesigns more and more difficult, and more expensive, all with less to show for it. Which brings us to the boxy gridlock we experienced in the 90s. Which is why standardistas get so angry about this stuff. We know that customers deserve better than this. We know that when customers find out how their designer painted them into a corner, it casts a shadow over all of us in those customers’ eyes.
The question that remains from all this is, how can the professionals in this field separate themselves from the amateurs? Really. I want suggestions. What concrete steps can we take to ensure that the good designers and developers, the ones who are always learning, who have a full and balanced skill set, don’t get lumped in (or worse, beaten out by) the ones who are locked in 1995? Who’s got an idea?
September 19th, 2008 at 16:06 UTC
If you can write a HTML/CSS best-practices rules validator that produces little to no noise (good sites getting bad raps, bad sites getting good raps), then you can have your own spider produce a best/worst of the internet. You collect core data, and then expose it in a fashion that people can play with weightings themselves (though you would obviously provide defaults)… weightings like which rule is worse, and how the ratio of number of pages in a site to the number of duplicated style informations, or amount of tag redundancy. Better yet, you could write a plugin (for Firefox) that mocks the bad websites and gives a medal to the good ones that someone is currently visiting — you might even go so far as to integrate it with Firebug, so you can have it jump to the categories of worst offenders of a website. Make it easy for people to fix their own problems.
We do the same thing in the world of C++ (PreFIX, PreFAST) and .NET code (FxCop), though the kind of integration I mention for the above isn’t as good as I might like.