In our environment, when a tester finds what appears to be a problem, they log it into the defect tracking system as an incident. We then have a daily review to classify these incidents and if it is so determined, it then becomes a "defect".
You can define a defect as a description of a specific problem identified during testing.
I believe the IEEE (www.ieee.org) has their definition (as well as other tems - such as "fault" - that are related to system reliability).
My definition of a defect is this:
Behavior of a system that doesn't match customer requirements.
Notice I didn't say "specifications" or Marketing/Engineering requirements. Now I acknowledge there can be a big debate over whether you should test to the requirements you are given to, or try to think beyond. From my perspective it depends greatly on project/product size as well as the maturity of the organization.
I try to remember that specifications are *NOT* perfect and testing to them may be what Engineering/Marketing want, but if the product doesn't meet the customer's needs, we'll all (eventually) be out of a job.
So hope this helps- please pardon any editorial comment that crept in.
In reviewing your post again, I thought the following might be helpful:
There can be many defects in a product- software, errors in marketing literature, poor labeling on packaging, wrong or outdated information on a customer support website, etc. Whether you want to log all those problems in a defect tracking system depends on whether you can get the people that are responsible for fixing such issues to use the tool. If they can't (or won't) you may very well still want to log the defect in the tool, but manage the defect outside the system (e.g., print out the defect and hand it to the person). Some defect tracking tools have, I believe, e-mail capability for this particular reason.
Or you may choose to just log those items that management truly cares about (probably software defects) and wants status on (e.g., how many severity 1 problems do we have to fix before the target release date?).
[This message has been edited by acquinas (edited 08-25-2001).]
QUALITY . . . is like buying oats. If you want nice, clean, fresh oats, you must pay a fair price.
However, if you can be satisfied with oats that have already been through the horse . . . that comes a little cheaper.