A few thoughts ..
(a) Your company will have produced ...
Most do, some don't .. it is good practice for the commander to know which sort of company it is for which he/she flies.
(b) If you are able (performance-wise) to follow the normal SID and the required climb profile it is advisable to fly that since ATC does not know your EOSID
Fair comment - but the ops eng folk need to have looked at this topic in detail to make sure that tracking is feasible from the approach and that net profiles, especially third segment acceleration and turn radius is OK. That is to say, not appropriate for the commander to figure out on the fly.
(c) To use a certain procedure you have to meet the required climb gradients
Do keep in mind that gradients are only part of the problem. Configuration changes, gross/net decrements, and turn gradient delta/radius complicate things.
(d) .. the commander may always decide differently ..
But do make sure you have a good story for the Enquiry. Be very wary of continuing visually in tiger country areas - the gradients are very shallow and not amenable to winging it as we go.
(e) You may get your engine failure at 3000' with the engine out procedure behind you... but will you make it above the 12000' peak ahead of you?
A good ops engineer will have looked at ALL the possibilities until you are above all relevant terrain. Unfortunately, not all ops engineers are operationally savvy and, if this is the case, the exercise should involve a joint ops eng/flight standards liaison.
(f) The point is that you only have to meet the gradient ..
Not adequate. Please do keep configuration changes and turn deltas/radius in mind.
(g) I would also commend looking at a topo chart if you'd be unable to meet the gradient
Topos give useful background - particularly in respect of the big picture stuff. They are not too bad at telling you that there is a BIG hill here or there ... however, they are NOT terribly accurate or useful when it comes to figuring critical net clearances.
Further the sorts of errors which creep into the topo world are legion.
In the real world of ops engineering we use the topo to figure out where the difficult bits might be and then, if it is really critical, a couple of fine chaps/chapesses draw the short straw and have a fun day out in the bush with a theodolite and associated kit.
(h) When visual a Mk I eyeball is also a useful tool.
So long as it is limited to staying away from big hills in the same way that we eyeball staying away from thunderstorms. Mk I eyeball is pretty useless for scraping over ridges and such like OEI.
(i) Are you aware that your gradient diminishes in turns
Something similar to operating at higher weights (load factor effect). The specific sums might vary a bit but, typically, expect something in the region of 0.6 - 1.0 percent decrement, give or take.
(j) It boils down to never go anywhere in an a/c that your brain hasn't been first.
That's not a bad philosophy ..
(k) The required gradient is an average over the procedure, so yes it may drop off in the turns
.. but how are you intending to allow for such things ? The main problem with the real world is that the pointy bits of the profile have a nasty habit of sticking through the gradient surface and hitting a hard bit ... averages give some guidance but don't hack it at close quarters.
Also, before we get too anxious, keep in mind that the gross/net decrement gives a fairly big delta the further we get away from the runway so it's not all gloom and doom .. but once the lawyers get a hold of it, they do tend to get introspective about black and white, rather than the real world of shades of grey.