There's an interesting discussion going on at the Association For Software Testing discussion list on non-functional testing. Basically, Danny Faught thinks the term is weak and is looking for alternatives. I happen to agree; here's my reply:
Danny Faught Wrote:
>Yes, "non-functional" is a rather bizarre term, though it seems to be gaining acceptance.
Yes, that's kind of a weird term. Personally, I have a problem with "requirements" in most commercial organizations. I have worked in organizations where the requirements process is followed to the letter, the documents are created, and someone comes over to the engineering group and asks "How long will this take to build?"
We reply "Well, as written, about two years. But if you take out bullet points C, F, and G, I can do it by myself in about a month." And, magically, there is a new version of the document with C, F, and G removed.(*)
If that was the case, could we ever really say that C, F, and G were "Required"?
Non-functional testing has the same problem. It is lingo that doesn't -quite- match up with common sense English. Now, I have given up on trying to correct the requirements term; it's just too entrenched. Now, then again eliminating non-functional ... we might have a shot at that.
I like Para-functional.
--Matthew Heusser
(*) - I am exaggerating for effect here.