Comment 35 for bug 882274

Richard Gaskin (rg4w) wrote :

There are many opinions about design, but one way to move beyond opinion to an implementation we can all have greater confidence in is to broaden the scope and frequency of usability testing.

It may be useful to note that the phrase "usability testing" appears in only two posts here, and unfortunately neither suggests doing more of it.

Earnest usability testing, employed in a way that sincerely seeks answers with an open mind and minimizes a priori bias, is indeed difficult. But it is also so very worthwhile, arguably necessary for something as important as Ubuntu.

In the 11.x releases we've seen published notes about only two or three usability tests. As useful as they were, they were done relatively late in the design process, obviating the opportunity to act on some of the issues identified.

Usability testing can be expensive, but one way to raise the return-on-investment for such testing is to consider adopting A/B prototyping early on, probably best done with paper prototypes in those early stages, long before a single line of code is written. User Interface Engineering and other excellent teams have published extensively about the methods and usefulness of paper prototyping, and it may even be the case that Jared Spool and his team at UIE might be willing to contribute consultation to such a high-profile project as this if asked.

There are also ways the community can contribute to increasing the scope of usability testing. This is probably best as a topic in itself, but in brief it should be possible to establish guidelines and templates, and identify members of the community with experience conducting such tests, so that valuable data can be gathered from a much wider variety of contexts and delivered to the design team at Canonical with close to zero cost to the company.

In summary, may I humbly suggest an increased role for usability testing as one way to move forward beyond debates centered around opinion, to unite the community around designs whose benefits can be quantified.