A general problem in GIScience is that there is too much data, from an arbitrary number of sources represented in a enourmous amount of languages. Therefore the solution is typically to use only one source of information and hard-link it to the map-rendering system. Such as Google Maps and similar.

A vivid discussion on NRKBeta’s pages (norwegian) started from their post on user’s comments on how they should develop a new system intended to be used by the norwegian trip association (typically hicking off road) for planning trips of any sort and be a portal for everything tour/trip/hicking related.

Not surprisingly a suggestion on using OpenStreetMap came along – spawning a “wild” discussion on the capabilities of open source maps/geoinformation and even open source in general.

I took a more lean approach to this and argued that combining “the best of every world” would be the best solution. And thus referencing to trends like service-orientation, loose coupling and similar. I believe this must be a quite good approach to the “problem” of having too much information, where subsets of different providers is appropriate to the task in question – regardless of cost, licensing and other more managerial issues. Atle posted an idea of making a similar approach in his master thesis. In response to this idea I sketched an architecture that could enable this, the architecture is inspired partly on the INSPIRE directive in addition to the general service-orientation.

Service oriented architecture for a geo-centric system

The architecture aims at illustrating where complexity of implementation may adhere. I believe however, that an implementation is feasible and possibly a rapid prototype feasible when taking a low-level approach while maintaining the conceptual idea. The numbered circles are to indicate reference points that is made in a wiki-page (norwegian..) regarding this architecture, the MS Visio file can also be found there.

What do you think of this architecture? Is it feasible? Is it overly complex? Is it good? Has anyone tried to implement something similar?

Several years ago I read a discussion on whether pirate music is an economic deficit or benefit for the music business as a whole. The argument were that the small, “indie” and non-label artist (and their music) got a significant increase in the market as well as the more commercial big labels artists didn’t decrease that much. So combined, they argued that the sum of the “tail” of the popularity graph, relative to the genre/labels/artists, was larger than the sum of the commercial labels, or the most popular music.

This fascinates me alot, and more recently I witnessed another exploitation of this “sum of tail” phenomenon on a FAST presentation during ITovation conference (video stream of fast presentation). FAST recognize that the sum of the “smaller” enterprises that needs enterprise search is larger than a single focused search strategy (public web search, i.e. Google, Yahoo etc).

Inspired by this combined with an interest of knowledge management, creation and acquisition, I thought of the way we acquire knowledge over a period of time. A study is a nice, illustrating, example for this. Over a defined period of time, say 5 years, we are in a process of acquiring a certain set of knowledge, for instance learning certain math skills, software development skills and similar. The set of intentional knowledge is pretty well defined, and we can assume that we learn most of it. However, in the same period of time we acquire other knowledge, often the result of a certain interest or more accidental. Knowledge motivated by special interest can be assumed to be a bit smaller in degree of specialization. Accidental knowledge may be a viril discussion, random web page, TV show and similar. Together this can be a graph similar to the one mentioned earlier.

Knowledge Acquisition over a period of time

So, the essential question. Is the sum of the “tail” of knowledge greater than the “body”? Seperation of tail and body is reasonable to put somewhere in the intersection of interest and accidental. I hypothesize that the tail is larger.

Yet, again, the title is somewhat misleading, well, but who cares:)

BMI, or Body Mass Index is quite a popular metric for defining what is “fit” (i.e. normal), unhealthy, morbidly obeses/slim etc. While mankind (almost) always have been obsessed by normality and especially fitness, the BMI may be a bit over-simplifying, but still, it’s gaining popularity. I’m going to let the discussion on whether it is a good metric or not be left aside, but personally find it incomprehensive.

Anyhow, being interested in formulas, and especially when quite complex formulas gain massive popularity (i.e. BMI), I came up with the idea of visualizing the BMI as a surface in 3 dimensions. I didn’t put a lot of effort into it, easy peasy Matlab crunching – and the result:

BMI surface in 3D. Height and weight at "ground" level (x/y axis)

I find the surface quite intriguing and immediately got the idea of taking this surface beyond the graphic visualization. How cool would it be to make a full scale version of this, for instance in some shiny metal, and place it in a right in the center of a “fashion capital”:) Then one could stroll along the coordinate system on the ground and look up to the sky to see there BMI. Artists out there – make this! :)

The inspiration may have come from the “bean” in Chicago (US), which I find very nice and it seems to be acting as a creativity center – look at the people acting like children, and how enjoyable it looks! (picture series at aftenposten.no)