The Ghetto of Web Design

I’m a graphic designer by training and the art director of a web design studio by profession. This presents a great number of frustrations on a weekly basis, chiefly a sort of professional self-loathing and a constant struggle as the ambassador of design in a hostile territory.

I meet a lot of web designers (and read rants by many more) that have debilitatingly titanic chips on their shoulders about their relationship to the design world. They imagine an ongoing clash between themselves and print designers. In this imaginary debate, print designers disrespect web designers for not subscribing to the same principles and beliefs about design. The web designer’s response is that “the web is complicated” and that print designers “don’t understand” the medium and shouldn’t “make judgments” about it. And all print designers make websites that look “poopy”. And they write “terrible javascript”. And they’re all “snobs”.

Now, its true that there are a lot of print designers who have hubrisitcally designed and built some awful websites. Its also true that there’s a lot of atrocious coding out there (but let’s not put that all on print designers). The sad thing about this argument is that its actually not happening at all, because print designers aren’t even engaged. Most designers who don’t self identify as “web designers” accept it as the rule that the internet has a serious dearth of good design. They don’t want to argue about it. They’d be thrilled if you could show them some good typography on the web, but no one’s holding their breath.

This isn’t a comment on the awesomeness (or lack thereof) of the web. It is the most profoundly world altering technology since radio and you’d be hard pressed to refute that claim. Its just that when people look at the web, its hard to get excited about the presentation.

The problem facing web design as a discipline is that its practitioners are unable or unwilling to judge the quality of their design work outside the context of the medium they’re working in. When we say that a website is well designed, implicitly we mean “as judged against other websites”. Its like telling a five year old she’s a great pianist after her recital. There are very very few great five year old pianists if you judge them musically, so you can only judge them against each other. When we say that virtually anything else is well designed, we mean that it is appealingly in concord with the visual culture we’ve been socialized to. Its thrown in the same mix with all aesthetic production in history as we understand it, judged according to our cultural relationship to the golden mean, academic painting, classic cars, pornography, video games, family photos and so on.

Web designers try to excuse their bad design with some silly claims about why their jobs are especially difficult. “The internet has technical restrictions” (as do all media) “The internet is dynamic and needs to change constantly” (newspapers? magazines?) “We can’t anticipate the content when we design the site” (ditto publications and signage systems). This last excuse actually brings us to the central problem with web designers’ thinking about their medium. Most contemporary web sites are designed as systems, not as actual sites. The content is abstracted from the design and always changing. The implication is that the design system is a thing that should anticipate and accommodate any arbitrary content that get’s put into it with no further design. In fact, web standards zealots (of which I am one in most regards) consider it a sin to style an element directly in the html document. This means that all styling must be specified separately from the content and more importantly that style should never be treated as unique or intrinsically related to specific content. This isn’t a limitation of technology, this is a philosophical belief espoused as gospel by many of the worlds most accomplished web designers.

Now, if you even consider applying this sort of thinking to any other medium it becomes quickly apparent how absurd it is. Imagine trying to design a magazine before the copy is written and expecting that no one will have to make adjustments after the fact, or making film titles without having the names of the cast or crew and expecting it to actually look good. Or designing a piece of electronic equipment but you won’t know how many buttons go on the thing.

I recently had to explain this to a colleague of mine and it clarified a lot of issues. We were discussing the impossibility of creating truly great typography in HTML. He argued that simply by setting letterspacing, weight, size, and line-height (variables controllable in html) he could achieve an acceptable typographic solution that would obviate the need for a hand kerned image. We looked at a few samples that used arbitrary text. I agreed that each sample had some element that was interesting, and certainly they were formally superior to unstyled type, but that they could never compare to a piece of hand set typography. Was it good enough? Sure. But in all of history, “good enough” has never been “great”, which means that if “good enough” is the dominant standard of web design, it is a medium destined for mediocrity.

We went back and forth. Eventually he ceded and told me that he was going to write a piece of javascript that would kern each letter, which started a whole other argument.

What this particular designer didn’t know (and why would he) is that he was describing a program that could dynamically generate kerning tables for a font. I explained to him “this is the thing that Matthew Carter has spent his entire life doing. He’s one of the greatest type designers alive and the best he could do was what you’re looking at right now (Georgia). Which is to say, despite being the world’s foremost expert in writing software that does this very thing, his astonishingly good product requires the finesse a of professional in order to get the most out of it.”

That’s when it struck me: The design of websites is based on constructing a set of rules, a program if you will, based on constraints and variables that are laid out by several disparate and infinitely more complex systems of rules (web browsers, css specifications, and fonts). The cumulative program presumes to anticipate and guide all possibilities of formal relationships based on arbitrary content of unknown value/meaning.

A font, by comparison is a complex program that specifies the formal relationships between about 300 known typographic forms that were specifically drawn to have familiar relationships to one another. There are a finite number of possible pairings of each of these characters (≈300!), but that number is so large that font designers only bother to specify the most commonly finicky relationships. Even with centuries of theory and knowledge about how to do this the best fonts require the eye and expertise of an professional designer to hand space the characters in order to get optimal results for a unique word form. Before the computer this wasn’t even the work of a designer. This was a specialized technical operation that was performed by people withe professional training.

Now, this is fairly technical and definitely nerdy, but if you’ve followed me this far you can probably see where I’m going. By comparison to a font, a web page is a limitlessly complex arrangement of letterforms and images that must coalesce into a cohesive and formally pleasing composition. Trying to devise a system/program by which we can specify or construct a reasonably pleasing outcome is a difficult task (one that keeps me employed). Designing one that results in a truly great composition, that considers the most sophisticated gestures and refines the finest details (the level of design that we expect from all other media) would be akin to creating an artificial intelligence.

To expand on the infinite monkey theorem: Even if you found super smart monkeys who could spell, it would take quite some time before they could type anything resembling Shakespeare. In design you could definitely come up with Joshua Davis, but you’d never get anything like Martin Venezky.

If web designers want to be taken seriously, if we are ever to be a legitimate design discipline, a few things have to happen:

  • Web designers have to accept judgment by the criteria of aesthetic paradigms, not technical paradigms. Web developers can concern themselves with other web sites, web designers need to concern themselves with DaVinci.
  • Web designers have to reject their dogmatic obsession with flexibility. Insisting that any piece of data or content must be ready to be presented in a different design context at the flip of a switch is needless, but more importantly it denies the possibility of presentation that addresses the unique qualities of the content itself.
  • Web designers need to educate themselves about the history of design. Too many web designers are painfully ignorant about any visual culture from before their awakening as designers. They focus on technical possibilities as opposed to aesthetic ones.
  • Web designers should think more about design details than tricks. In all the years I’ve been a web designer I’ve read countless articles about using javascript to achieve such and such affect (“without the use of images!” barf.) and maybe only a handful that even suggest that designers should consider line-spacing.

There are probably a hundred other things I think web designers should do and don’t, but I’ll leave it at that for now. Feel free to leave your own suggestions below.

Note: May 17, 2009
I recently discussed this topic with a designer whom I have a great deal of respect for and he accurately pointed out that this article treats design solely as an aesthetic practice. I would like to clarify: commonly there is a distinction between “graphic design” and “web design”, graphic design being the domain of printed media and web design being related to everything in a web browser, both the graphic systems and the technical systems involved in production. There is no common term that designates the graphic design of the web, which is the subject of my criticism. I think the lack of semantic designation for this important aspect of the medium further supports my stance.