Chaps,
I have no interest in arguing semantics - a fairly pointless exercise unless you are into editing dictionaries.
As a performance engineer I use and am comfortable with the philosophy of gross/net gradients - if you are more comfortable with something else in the way of descriptors, then that is fine. One of the other correspondents suggested the use of 152 ft/nm in lieu of 2.5 percent - all fine and dandy - one should just use the slant with which one is most comfortable. The main aim is not to go hairing off with no idea of the situational awareness requirements for the particular runway and aerodrome.
I think that we each understand where the other is coming from.
More importantly, I suggest for the discussion, the average pilot has a limited, if much at all, knowledge of the ICAO Doc and Procedure Design but is fairly au fait with the idea of the heavy aircraft segment gradient system.
The point remains, though, that the run of the mill ICAO (Pans Ops) SID, without annotation, involves a 2.5/3.3 gradient profile and this is what the average civil pilot has to content with.
An ops engineer is not going to use the SID track inferred obstacle profile without checking the discrete obstacle data. This is due to the fact that procedures designers generally work their obstruction profiles a little more conservatively (due to the terrain data with which they often have to work) than the ops engineer can tolerate.
However, for the pilot left to his own devices, the SID figures at least give some information and this was, I suspect, the intent of the original question.
Most importantly, at commercial weights, it is potentially suicidal to head off along the SID following an early failure.
Further, if the operator doesn't consider the failure following V1 set of cases (and I believe that many don't) then the only reasonable way to address the failure case for the conservative pilot is to follow the nominated V1 failure escape path (if provided) for all takeoffs.
If a specific escape is not provided, then that may permit an inference that the straight ahead path is OK ... but for how far out ? ... best ask the guys who did the analysis.
[ 02 September 2001: Message edited by: john_tullamarine ]