(Manas Garg, who had organized the recent POCC meeting on Usability, was inspired to write this article after attending that session.)

Ultimately, we build all the systems so that they are used at one point of time, and the system that we build will be used only if they are usable. If a system is not usable, how will someone use it?

It is very common to come across doors that are supposed to be pushed to open (and there is a sticker close to the handle bar which says PUSH) but most of the people will pull it. Also, it is very common to see doors that are supposed to be pulled but people end up pushing them. It’s not really a problem with the people. It’s a problem with the design of the door. It’s a usability issue. There are doors that are pushed when they are expected to be pushed. It’s all in the design, isn’t it?

When Gmail was launched, it was an instant hit. Even though it didn’t do as many things as other email services did at that point of time (like support for all browsers, drafts etc) but it was still a revolution by itself. And the reason was simple – usability. What was the need of Gnome project? Wasn’t it the usability of GNU/Linux for non-programmers?

Two levels of system usability

1. Conceptual usability

A system is not an island. It is connected with various other entities by various means. And it shares one relationship or the other with those entities. A conceptual model effectively captures these various entities and what kind of relationship our system has with these external entities.

The model of Britanica Encyclopedia is that a group of experts together create an encyclopedia over important topics which can then be read by people. The model of Wikipedia is that people (and that includes everyone in the world) can help in writing the encyclopedia that everyone can read. The model of Google Knol is that experts can write articles on specific subjects which everyone can read. The readers can suggest improvements that the original authors can incorporate.

The conceptual model must be made usable. The entire system will eventually be built on top of the conceptual model.

2. Interface usability

So, we have settled that a system maintains some relationships with some external entities. This relationship is exposed through interfaces and we need to think through the usability of those interfaces.

Gmail is an excellent example of interface usability. As I mentioned earlier, there were several email services when Gmail was introduced but its interface usability was far superior. This is an example of how the same conceptual model can be presented to the user with completely different interfaces.

The conceptual usability is a must to help the user understand the system. Wikipedia is an encyclopedia which can be read and modified by anyone. And interface usability is a must to help the user *do* something with the system. On each page of Wikipedia, everywhere your eye goes, you’ll find a link to edit the page or a section thereof. It almost invites you to modify the page.

Methodology and Evaluation

There are ways/methodologies to design usable systems. The usable systems are still created by people whose thought process naturally evaluates the usability at every step. However, these methods can make the system designer better grounded in the real world and make him/her more efficient too.

Much like a system cannot be declared as functional till it is tested for the same, usability of the system has to be tested with equal vigor. After all, if the system is not usable, who is going to use it (even if it is functional)? And there is only one way to test the usability of the system – by letting target users use it and that too without providing any guidance. Closely observing the users can be an eye opener many times.

And the system designer/developer cannot verify the system usability. In the course of developing, the developer has learnt too much about the system and knows exactly how it works. However, the user would not have so much of knowledge about the system and may not attempt to do things in the same ways. If you don’t believe me, go and re-read the doors example.

And what else?

And first thing is still the last thing. We need to create usable systems because nobody uses unusable systems. And very few systems are usable by accident. Most of the usable systems are developed with usability as a focus.

Further reading

• Overview of Usability from Jakob Nielsen
• Wikipedia entry of Usability
• US Government guidelines on usability
• Basic usability guidelines from Yale
• POCC meeting on Usability – PuneTech liveblog
• Aman King’s article on Usability event – Geek Night at ThoughtWorks