Thursday, February 17, 2005

A piece of IA pie: little, micro, lite or guerrilla?

Instead of Big IA (excuse me: UX) this time a look at the smaller pieces of the pie.

First we had "little IA" versus "Big IA" from Peter Morville's column "Big Architect, Little Architect" back in 2000. Then last month, implicitly we had to to adopt "micro IA" as opposed to "macro IA" because that's Andrew Dillon's new term for Big IA. And now there's "IA lite" (vs. "IA classic"?), Dave Rogers' nickname for the IA that a webdesigner could learn easily, according to Joshua Kaufman.

The first two adjectives ("little" and "micro") refer to the same activities: deep IA work focussing on structures, search, semantics, metadata, ontologies, taxonomies, or controlled vocabularies.

The third one ("lite") is new, and well chosen I think. I shudder when I think of Lite Beer and similarly so when I think of Lite IA or IA Lite. But part of that is snobbery: Lite Beer has its place in the spectrum of beers and so may IA Lite find a place as well. As I said in "Big IA is now UX":
Related fields, placed in the horizontal line of the T-model, have interests that overlap with our shallow subjects.
IA Lite focusses on these shallow subjects, like navigation, labeling, and layout.

Maybe IA Lite can morph into the IA version of what Jakob Nielsen called Guerrilla Usability:
[..] simplified usability methods [that] can be a way for a company to gradually build up its reliance on systematic usability methods, starting with the bare minimum and gradually progressing to a more refined lifecycle approach.
It's a shame that Guerrilla IA seems to be taken by Lou Rosenfeld to mean something else, otherwise we could use the same words that Jakob used and define Guerrilla IA as:
simplified IA methods that can be a way for a company to gradually build up its reliance on systematic IA methods, starting with the bare minimum and gradually progressing to a more refined lifecycle approach.
Who knows, if IA Lite catches on, maybe one day IA will fall pray to what fellow Design Engaged participant Tom Coates called
"The Mass Amateurisation of Everything"
and we have the IA processes, methods and tools in place so that everybody can be an IA.


Anonymous Anonymous said...

We already have IA Lite - it's called card sorting...

10:58 PM, February 23, 2005  
Blogger beep said...

Card sorting by itself isn't enough to qualify as a set of lightweight methods, both by definition and if you look at the areas covered.

Card sorting can be seen as a method combining user research (inventory of preferred terms) with structural design (use groupings as final structure) but only if (a) you assume you know the full content and functionality beforehand and (b) you can find an easy way to calculate the average of the exercise's end result to use as the structure.

At the very least, card sorting will have to be complemented with lightweight methods for content inventory, task analysis, evaluation and documentation of the designed structure to qualify as IA Lite.

9:49 AM, February 24, 2005  
Anonymous Anonymous said...


I am baffled by the need for distinction; big/little, micro/macro, lite/heavy? There is such strong connection between these, I do not see having one without the other or a project that would require one without the other. Such schizophrenia seems unlikely to contribute to the acceptance or development of IA. Do we need to bifurcate the discipline?

7:29 PM, September 16, 2005  
Blogger beep said...


Do you own a car? Is it "the" car, or a particular one? Why? One can choose a familiy car if transporting small groups of people and their luggage is the goal. A citybus does the same, is cheaper for one-time use, but you wouldn't want to own one. And both would look out-of-place on a race track.

The same is true for IA activities; some can be done in a hurry by an (empathic) untrained designer, others require years of practice, explicit domain knowledge, and/or business insights. Some IAs have the luxury to spend weeks on card sort exercises, other have to come up with a tested structure in a week.

Yes, the average project requires an average amount of IA activities, but tell me: when was the last time you did an average project? And was it the same as my average project?

That's why I describe different approaches, foci, viewpoints; to come up with a model that matches everyone's needs, that scales from small to big, from slow to quick, from deep to shallow. And yes, they are all connceted, but they also have their differences. I like to put labels on those differences, these attributes, to allow us to compare and learn.

3:35 PM, September 18, 2005  

Post a Comment (moderated)

<< Home