Humane Information Architecture

This article is subtitled "An interim definition" because I am pretty sure I can't come up with a definition as wonderful as Jef Raskin's the first time around :)

By sheer plagiarism, I will start with "An information architecture is humane if it is responsive to human needs and considerate of human frailties." Nothing wrong with that one at all. But it could stand a little clarification. Which humans? Which needs? What frailties?

Are we talking about just the end users here? Does the humane information architecture need to be responsive to the needs of the IAs working on the project?

What constitutes a need from an IA perspective? To me, this will be a two-part answer:

* whatever the requirements documentation says is a need, and
* whatever the IA believes is a need and can be reasonably negotiated into the requirements set that is vital for the successful operation of the system from an IA perspective.

Examples of the latter are legion. It is not as if the IA is the Knight in Shining Armour come to rescue the poor end users from the evil Project Manager - far from it - but there are times when things get missed in the best of requirements sets. So we could take "do the best you can within the allowed budget in a humane way" and add to it "to meet the needs of all users and in so doing, make the system both useful and useable".

And frailties? There are the normal interface design motherhood statements like:

* "there should be no unexpected behaviour",
* "clickable regions should be readily identifiable",
* "use visual hierarchy to ensure that the most important stuff is easily found",
* "the navigation system must be clear and intuitive", and
* "use navigation labels that make sense to your audience".

How do you find out if these are enough? By testing - informal if needs be, but test early and often, with a minimum representative sample of ALL user types.

So now we have this: "do the best you can within the allowed budget in a humane way to meet the needs of all users and in so doing, make the system both useful and useable, and test to ensure that it recognises and compensates for apparent human frailties". This is a bit of a mouthful - I am sure I can come up with something better in time!