There Is No Cat - How to make software work

Sunday, March 30, 2003

How to make software work

Dorothea Salo discusses the cult of the programmer, an insular world where programmers never ever ever talk to the people who are supposed to use their programs before they create them. She comes to the remarkable insight that if you talk to the users, you wind up with better software.

<sigh> Usability people have been screaming about this for years. There are entire professions dedicated to studying users and communicating with them to help programmers create better software. Ethnographers and anthropologists have found new uses for their esoteric fields in corporate settings, sussing out the needs of users. Techniques such as contextual inquiry allow development teams to glean insights into the way users use their tools before programmers start programming. Usability tests help refocus efforts once programming has started by checking to see if the assumptions programmers have baked into their programs match the situation on the ground.

I don’t mean to rag on Dorothea, but given all the ink expended on usability in recent years, I’m surprised she’s not aware of all this. I know usability is a dirty word with a lot of people, but there are facets of the profession that really do provide a decent response to the problem Dorothea points out. Sadly, it’s a response more often honored in the breach than the practice, often brought into the development process at the last minute to sprinkle usability dust over a product that’s already pretty well fixed in form if it’s used at all, rather than being involved from the beginning in shaping the direction of development.

Posted at 1:28 PM

Comments

I do know, actually. But there’s a difference between studying a job and *doing* it.

I think programmers ought to get out there and *do* jobs, not just study them. And that’s the difference I was gunning for in my post.

I don’t doubt the techniques you mention are good proxies. But proxies only get you so far.

Nobody’s tried direct experience, that I know of. Wish somebody would, is all.

Posted by Dorothea Salo at 2:02 PM, March 30, 2003 [Link]

I guess I’m just dense, because I didn’t quite understand from your original posting that you wanted the programmers to actually *do* the jobs. While I don’t doubt that that would be an improvement to the situation you describe, I think the understanding gained by doing that would be pretty superficial. The insights that can be gained by doing a job for a few hours are quite different from the insights that can be gained by doing a job for a few years, and those insights are the kind best teased out by the techniques I mentioned. So I think it would probably be a useful exercise, but probably not as useful as the techniques used by usability professionals.

Posted by ralph at 6:12 PM, March 30, 2003 [Link]