Actually, weld termination is very often not done correctly. Many times a welder will wrap the weld around the ends, such as on a shear tab plate on a column. However the weld symbols usually show a 2 sided fillet weld but not the all-around symbol. So in this case, wrapping the ends is incorrect. If the drawing did not show to hold back on the ends then not bringing the welds right to the ends is also incorrect.
Since we are all creatures of habit, you will usually see the welds wrapped because A) that's the way we've always done it, B) nobody told us any different, C) I just weld the way the foreman tells me to, and D) it's easier to make the weld look good that way.
Most of the time the customer doesn't object even though it's wrong so the wrong lesson is reinforced.
This is kind of like square-butt weld symbols without a specified weld size being taken for PJP or even seal welds when, technically, they should be a CJP weld per AWS standards. That problem is compounded by detailers who use the unspecified square-butt when they intend a seal weld. After that happens often enough, welders, foreman, and company presidents will jump up and down saying you are wrong when you try to tell them it's really a CJP weld. Especially when a call to the detailer verifies that all he intended was a seal weld so now who's the dummy?
But all that's all in a day's work. Sorry for getting off the subject a bit.
Chet Guilford
In the 2002 code all this has been revised to Para.2.8.3. Are you bound to the 2000 code on the job that this Inspector is checking? What exactly does he have the problem with? Is the profile not full enough for the effective length required? or are you talking about the end returns? End returns usually are twice the fillet weld size, with a max. of 4 times the weld size. (ie. 1/4" fillet would return for a 1/2", 5/16" fillet would return for a 5/8",etc...) 
Forum
