10 Aug 2005 Hight of Un-usability or The Front-end for a Front-end
This is a story that makes me sigh with both relief and grief.

A couple of years ago we bought (what we thought) a state of the art XYZ software, from the no.1 vendor for that domain Zoofrel Inc., for internal use. (real names changed to protect their identity and reputation). We paid tonnes of dollars for that (may be several hundred for each user). When the attractively marketed and nicely wrapped box arrived, we were very happy to replace our decades old software with the new one. It was a typical application software - a database at the backend and a web based user interface as front-end.

On the day of its deployment, most people got disappointed! While it met our functional requirements to a good extent, and had a really good performance and scalability, people had really bad user experience. The user interface was so badly over engineered that no one could make real use of it without having to pull down a dozen menus and clicking a dozen buttons. The tool literally hid the information that was in the database backend. I don't know who to blame - either the developers at Zoofrel Inc., or the folks who assembled it for us.

I realized that instead of spending my precious time futzing with menues and buttons I can save myself some time writing a 100 line perl script to query the database and list things I would often want to see. [many thought that my pathetically slow and hurriedly written 100-line perl script, saved the day for Sun, and I got a "People's choice award" from Glenn Weinberg (VP of Operating Platforms Group), that year].

We couldn't throw away that software, we had paid lots of money and hired people to integrate that for us. That is like eating an expensive but fowl and rotten dessert, since you paid for it. While most people now really hate Zoofrel Inc., and we would never buy another piece of software from them, we had to live with what ever we bought. Life perhaps made a little easier with my perl script that displayed what I needed. However since we had no direct interface to modify the backend, we had to still use the unusable web interface.

Recently Venky wrote another perl script that actually automatically logged in to the Zoofrel Inc's horribly bad web-based front-end, automatically clicked all the required menus and buttons and successfully posted updates to the back-end database.

What he wrote infact is a front-end for a front-end. A really simple and usable front-end that uses the real front-end as a proxy to update the databases. Finally we were all so happy! We don't have to use the ugly and unusable tool! Venky got a "People's choice award" this year!

Moral of the story: usability is far more important than most people think. Usability applies to everything - APIs, command line parameters, configurations or programming languages, not just for GUIs. Usable software makes the people love it, unusable software makes people hate the vendors. For e.g., Google's main asset is usability, there is hardly any other serious technology.

Thinking from a OS and software vendor perspective, many of Sun products might have actually suffered the fate of Zoofrel Inc's software. Our customers could have written their own scripts on top of our software to do their daily jobs, or might have just rejected our software since they found it unusable (it is more easy to throw a rotten free cake than an expensive one!). If you had come across such unusable Solaris software please speak - we are listening. We have a team for Approachability or Keep It Simple Solaris.

Tags: Usability OpenSolaris Solaris

Link | Comments [4]