Main Points

Question: Where does a design come from? It may be the hardest question to answer of the entire design process. Many designers, but especially beginners who are new to the process, feel more apprehension at the opening stage of creation than at any other time. It doesn't have to be like that; design like anything else has a beginning and a process that you follow. Don't think... just do, then you'll be fine. Answer: hard work and research. Why does the owner want / need a website? Do they, really? What is the site's function? What measurable goals will define if the website is successful or not? What are the usability limitations imposed on this user interface (UI)? What are the accessibility issues relevant to creating the UI so that the website makes the best use of universal design principles? This can be intimidating for people; there's a lot to think about. The beautiful thing, however, is that each of these many questions has an answer, and that answer is an important piece of the puzzle that is your design. The more questions, the more pieces and soon you're on your way. So let's go.

Concept Creation Process

This process was not so difficult as it may be for some sites. Earth Chronicle is at heart a history website. It is a public domain library of digital texts. Taking these two concepts as a starting point, the idea of doing something along the lines of a book came pretty quickly. But that doesn't minimize the process. We started with the topic and purpose of our website, then started considering the limitations (real world references, usability, accessibility, etc.) and ideas generated by them; form follows function. So we started considering the function of the website and basically it's a library. Old books have always held a special attraction for me. The hundred year old Latin texts I bought from Powell's Books in Portland, Oregon while attending Portland St. The illuminated manuscripts I've read from Medieval Europe either in books or museum displays. As I played with that idea however, it became hard to conceptualize how a book would really fit any reasonable design.

The "Old Book" Design

Presenting a book might be easy onscreen, but figuring out how to present content within that design was immediately daunting. Using a fully opened book, while gorgeous, instantly posed the problem of how to deal with the trough between the pages, called the gutter in printing terminology. Having a massive dark line running down the middle of the screen was not attractive or functional. I also considered attempting a "one page" design where the book would be open, but the gutter would occur on the extreme margin of the design, as if you were reading one page. However, the concept felt fundamentally lacking in unity, there is a missing page in the design and it's just off screen where you can never see it. This was not necessarily a major problem; I jotted this down as a possibility, but it was not one I especially liked. I kept looking for other avenues. In the various permutations I considered, one I liked better is pictured here. Here the gutter serves as the visual division between the content area and the internal page navigation. Something the gutter does provide is an extra area of white space to either side of the line. I frequently struggle to find white space in my design, especially for content rich pages. You'll notice I also thought of the idea of adding drop caps like an illuminated manuscript. Not a major idea, or anything brilliant, but one of those nice touches that help pull a design together. It also, by it's malproportions serves the typical purpose of all drop caps, helping the reader find the beginning of the text through emphasis.

A hand-drawn sketch of the "Old Book" design. 1. Old Book design

A book design also presented another problem, While an open book has pretty reasonable length and width onscreen, I prefer the behavior of fluid layouts. In a fluid layout, this would look good, only if the amount of content kept the bottom of the page onscreen, i.e. extremely small pages would look fine, but on a page with more content you would start scrolling to read the rest of the page. When the page length seemed to portray a book 10x as tall as it was wide... it would start to seem like a very strangely proportioned book indeed. It's possible to work around this by making the words scroll on a fixed page, but that's a very unnatural behavior to see in a book. It's also counterintuitive and therefore bad from a user interface standpoint; not to mention I've never seen a fixed element that really behaved. If a user scrolls quickly, the browser seethes with turmoil, and the fixed elements bounce up and down in a herky-jerky motion. It corrects when you stop scrolling, but all these points were the design liabilities the “Old Book” faced and I was looking for alternatives.

The "Ancient Scroll" Design

A hand-drawn sketch of the "Ancient Scroll" design. 2. Ancient Scroll design

This is when I realized that books, while certainly old are not the only formats for communicating information before the advent of the internet. In fact, books are a relatively recent invention in most cultures. The three earliest forms of writing material were the clay tablets of the ancient Sumerians, the papyrus scrolls of the ancient Egyptians, and the bamboo scrolls of the ancient Chinese. Clay tablets would be clunky looking and unattractive and bamboo scrolls are too busy looking. They form a series of stripes light bands of bamboo with writing and almost equally wide bands of space in between them. These gaps are created by the pieces of multi-colored string which tie the bamboo strips together and create a colorful grid-like effect that would be much too distracting for a web design. A papyrus scroll on the other hand immediately presented possibilities. At first, I envisioned the scroll looking fairly traditional, a vertical tube shape of rolled paper on the left, paired with another on the right. The text of the website would be written between these two cylinders. However, this left me with the same basic book "shape" and the same basic problems as before. A fixed layout would look good, but be jumpy when scrolled and the text would appear to move unnaturally on the page. A fluid layout would look out of proportion except for pages with an extremely small amount of content. It's interesting, it's different; but it's a new dance to the same old tune.

The "Medieval Scroll" Design

A hand-drawn sketch of the "Medieval scroll" design. 3. Medieval Scroll design

What solved the problem for me was research and considering the functions and real world equivalents of what I was building. While a web design is exciting because it has different limitations and can break old boundaries, a design is there to enhance not detract from content. A design that's too radical is a distraction; I wanted my electronic painting to feel organic and familiar. So, I naturally started thinking about how the scroll works. The primary action of a scroll is when you unroll it to read it's content. The Ancient Scroll design is problematic because it moves horizontally and horizontal scrolling doesn't work well on a computer. That was my problem. Onscreen, people can scroll vertically very easy, but horizontal scrolling is a !@#$^#@$. It was a short hop to useless wishing that the scroll could move vertically, but the Ancient scroll was turned the wrong direction. That led to the not so useless thought that “Who am I hurting if I did that? What if I flipped the whole design?” That's what broke the user interface open for me.

While not as old, the image of a Medieval or Disney / fantasy scroll has become a common and well known shape that's comfortably familiar to many people. This scroll is not wound around a pair of vertical spindles, but simply rolled horizontally into a curved tube shape. Ideally, it's tied with a piece of ribbon or string to keep it "closed" until the knot is untied and the scroll is unrolled with the horizontal tube shape of curled paper at top and bottom. This layout would conform to the shape and length of the content on the page naturally. A fluid layout would not seem unusual if extensive content made the scroll appear ten times as long as it's width; the design looks natural with any length you want. In fact, comedies frequently employ the the Christmas list (or list of complaints) which looks normal by width but when the end is unrolled, it hits the ground and rolls across the floor. The Medieval scroll also lends itself to textures that were more visually useful to me. I wanted to give the page a realistic aged look, but true papyrus has a very obtrusive texture; the grain consists of heavy fibers and intermediate tissue that turn different colors as it ages, exaggerating the already heavy grain. Medieval scrolls have an old fashioned feel, but more modern surfaces. Paper or parchment are less visually distracting while still distinctively aged and authentic.

You can see how I derived this. I started with the purpose of the site and considered how it functions. I used these ideas to work back and forth between site requirements and real world studies that could serve as a basis for the design. I used these mental models and the principles of usability and accessibility to narrow and guide me towards finding a more functional, more interesting, and more attractive design. It's a process of combining the fruits of research and experience with challenges and limitations; the liabilities of your models are just as, if not more informative, as their strengths in leading you in the right direction. For me, design is literally the act of putting a puzzle together, where the first step is having to hunt down and find all the pieces. Each piece reveals a little more of the picture and gives you clues as to where to find more pieces. Once they're all together, that's when you see how it works and you can start to put the pieces of the puzzle into place.

Developing the UI Design Plan

Now we have the basics of an attractive functional layout in mind. Now its time to take stock of the strengths and liabilities that we'll face in putting together this design, to shape exactly how we'll approach the user interface. For one thing, in order to get the page curls in, I need to present the entire scroll. I don't have to, I could just present the writing surface itself. However, if I do that means dealing with the gap between the edges of the scroll and the edges of the screen. This already makes me think of where the scroll would be. If it hasn't become obvious already, I use real world references extensively in my work. Partly, this is just how I think, but it also blends well with my strengths as a designer; therefore, rather than working around and modifying my mental design prejudices, I've honed and developed them as part of my personal design process. Exactly how you approach this part of the project will depend on what mental strengths you have and how they jive with your strengths of execution. My skills in image editing and compositing dove tail beautifully with distinctly realistic models, and the fact that I personally like this kind of artwork is how my design process all comes together.

So for me, putting the scroll in it's environment is important background to understand the look. I think of it on a desk, maybe with a quill standing in an inkwell nearby. Books are piled up around it, some scrolls sorted out into the pigeonholes at the back of the desk, perhaps even a parcel of folded and wax-sealed letters tied up in a bundle. A lantern near the back left corner of the desk provides the soft light this evening now that the sun has gone down, though I can imagine it's light is still visible in the West. Or perhaps a candle in an old heavy brass candlestick with a scrollwork handle. The light throws a soft warm glow across the paper, as an evening breeze comes up the hill bringing the sound of crickets and sighing grasses from the fields and forests beyond. While the design needs to focus on the scroll to provide maximum readable space, it's important for me as a designer to determine it's context. At the very least, this would put a table background around the border and maybe some other items can sneak in too.

Fluid layouts are definitely trickier than a fixed-width layout. CSS is certainly better designed for fixed-width work, but over the last couple years it's no longer a truly professional option; the problem is that browser behavior is fundamentally fluid. With the explosion of high definition and mobile devices, users are now routinely resizing their browsers from a couple hundred pixels wide in a palm computer to maximizing it in a 2000+ pixel wide HD monitor. There is no option to build a fixed-width UI when screen widths vary across an order of magnitude. (a factor of 10:1) I've always preferred fluid design for usability, and facing these realities in 2009 I never considered doing anything other than a fluid layout. But that brings some challenges.

I'm basically trying to put a custom image border around the page content. Any surrounding border runs into the following difficulties. First, the corners have to be their own images, much like applying rounded corners to any element on the page. You need the top / bottom and side borders to repeat (x for top / bottom, and y for sides) so that the border will match any height or width; this is basically the magic that makes the layout fluid. This means that the corners have to be disconnected because you don't want them to repeat. This makes a total of eight images, top, bottom, left, and right sides, plus the four corners. This is a daunting task in CSS since you can't apply multiple images to an element, only one. So applying eight images in a fluid layout requires coordinating eight separate HTML elements in perfect fluid concert. Using images to create a fluid border is not for the faint of heart. This is a large element of risk in building this layout.

With a fluid layout, you also don't know how wide or how high the page will be. I can design the repeating borders to align with one of the corners using CSS, but the other corner will be an unknown and changing distance away depending on the users browser configuration. This limits some of what we can do. I'd originally thought of this design as having paper with a very rough border. This had to die what was for me a very painful death, since the border has to merge at any point with the opposite corner. Handling these image transitions may be another key area of risk. While it may be technically possible to get the image border working, getting the images to look good at all points may be a whole other story.

We may be able to fudge this a little. If you have a one pixel wide detail it will display fine or be hidden behind the corner image, these are free and you can have as many as you want. If you have a two pixel wide detail it could be split in half; that could be OK if the remaining detail is identical or similar enough that it can masquerade as one of your one-pixel details, etc. A second technique is to add an image as an overlay. An image that you position so that it won't interface with the corners. This can be used not only to achieve edge effects, but positioned in the content section as well. In this layout, it's possible to insert wormholes or burn marks on the page.

When you actually start putting together a layout there are all kinds of issues that you have to work around many of which don't appear until you're in the middle of implementation. However, these are the trouble spots and ideas that came up as I was working out my sketches. The rest of the design will come together as part of the process. But you don't want to be running up against insoluble problems when you're finishing a design. While I'll be working the design up systematically, expect these risk factors to be front-loaded as far forward as I can reasonably get away with. Starting known risk factors early gives you more time to solve them, and in the worst case scenario... that you can't, gives you more time to reformulate your user interface.