Any scripter relying on the linking order of an object to find out the link number of a child prim is so poor a scripter that they don't deserve being taken into account, since any modification to their scripted object (removal or addition of a child prim, for example) would break their poorly written script. I use heavily ll(S/G)etLinkPrimitivePararams() in my scripts, but I never make any assumption on the link prim number and never hardcode them in my scripts, instead using prim names to find out what is their actual number.
I'm sorry, but I'm not going to make the life of poor scripters easier: they have got to learn scripting properly !