And… yet another web app. A POC, or MVP, pivot, all this biz speak. Idea (“hypothesis”) was to do a catalog, user generated content (UGC), semi-structured (folksonomies, but with added semantics), which can then be drilled down, interactively, in un/foreseeable ways. Nu, I had some ideas.

Anyway, Paula Borowska explains MVP goals/principles well:


  1. Test product hypothesis
  2. Gather feedback
  3. Least effort/cost
  4. Viable, not mere mockup (realistic test has validity)

Or, as extreme programmers (XP) put it, decades ago: “do the simplest thing that could possibly work”. ;o)


  1. Items are essentially specsheets, detailed descriptions of real-world stuff, like, say, (models of) robots.
  2. Specsheets just list data points. Eg, “name: Asimo”, “color: indigo”. And here’s where folksonomies come into play: points are UGC, too.
  3. Classes of points, to be precise. Eg: size, model number, battery life, color, Pantone color.

Point definition

Semantic layering made possible by structuring point classes, eg:

  1. Title (unique string)
  2. Single or multiple values, eg for list of alternatives.
  3. Arity: number of dimensions (not necessarily of same type). Unary for simple values. Eg, “size” could be ternary, for LxWxH.
  4. Type: number, text, color, Boolean, price, image, URL, attachment, person, geolocation…
  5. Discrete/continuous
  6. Units: length, ratio, mass, surface, volume, pressure, force, time span, timestamp, color, yes/no, on/off, long+lat, vocabulary/any text…
  7. Precision



  1. Specsheet: table, presumably?
  2. Point definition: form, obviously?
  3. Search results: ah, here the fun begins. Eg, when filtering the catalog by a quantitative point — plot the results.
  4. Interactive: my primary hypothesis is that results views should enable further drilling through interaction, such as selecting ranges (with pointer: mouse or touch) in a graph.
  5. Inspiration?

So, now… to prove it works!

Comments are closed.