Saffron can be described as a multidisciplinary (HCI+NLP+DM) effort to build a system enabling easy and meaningful exploration of scientific literature. Each researcher/practitioner knows how hard it is to find relevant bits of information.
Below the surface, however, Saffron is a challenge. I will focus on the HCI part.
The challenge is to simplify access to complex data and support complex user needs. The data is complex since it contains several resources of equal importance, such as publications/authors/topics, which are highly interconnected. The user needs are complex since they are diversified: from general (i.e., exploration and getting insights into a research field/organization) to very specific (i.e., finding specific people or publications).
Simplification aims at decreasing perceptual, cognitive and motor burden. The hard design decisions include prioritizing and limiting supported scenarios, restricting what to show, how much to show, in which order, balancing visual complexity vs potential benefits, choosing effective (and minimalistic) way of communicating information structure, assuring uniform interaction model, assuring experience of seamless exploration, and giving away a lot in favour of simplicity and straightforwardness.
Simple things are hard to come up with. Too often complex data or complex user needs are matched with equally complex UI. Giving away is hard since it is difficult to perceive subtraction as addition.
Saffron is also interesting since it marries HCI with NLP. And I believe both have unexplored potential when considered together.
Saffron is work in progress. We are aware of some issues but we appreciate all kinds of feedback.
I decided to include some design sketches from Saffron. In the first group there are the initial digital sketches that were created quickly, with little refinement, using wide strokes, to try and see various layout concepts and interaction sequences, and to be disposed without (too much) weeping. Further we have a group of sketches, made with a pencil, which are more refined, but still easy and quick to do. Only better ideas get here. Look closely at some techniques of approximating details quickly and reliably – that is, when you screw up your eyes you get the accurate big picture. By the way, I also value more refined sketches at this point. Sometimes it is the details, styling, or various visual elements that trigger the ‘Aaahhhaaa!’. The last group of the pictures are not sketches. Only few final ideas deserve this level of refinement (finally this is about your time). They are made in a graphics program, not in the CSS/HTML or a GUI toolkit. The reason is that in a graphics program you can create the views faster, you can also explore various alternatives faster, and you have more tools at hand to assist you.
Sketching is a way to find and answer some of the questions quickly, the questions that would be otherwise found too late. There is of course more into sketching but this page is about Saffron really, so let me step back now.
