Tom, I am so confused. I feel like we're saying the same thing. I'm tripping over my wording, and maybe that's where all this confusion is coming from.
If it's in the TCDS, that is our basis for installation*; I worded that poorly, I did not mean basis for approval, since that part was already done
. As you said, logbook entry only. (*yeah I know, it's not an industry standard term, again, tripping over wording)
The IPC is, as mentioned prior, used as a part number reference and assembly reference. The maintenance manual is acceptable data, the structural repair data is approved data (if it has the proper signatures and stamps), but the IPC has always been something extra. BUT, as an example, if the mx manual references something in the IPC, then that reference becomes acceptable too. Same with approved data making references to the IPC.
Couple references stating the IPC isn't approved nor acceptable:
One reference (bottom of page 48)
Second reference (question 18)
This was also in my jeppesen book.
As a sidenote, you're asking me to prove the IPC not approved or acceptable. Proving a negative is extremely difficult to do (one of the reasons this whole thread has gone crazy; it's really damn hard to prove negatives). Rather, it would be a lot easier to find references stating what IS acceptable/approved data and make an inference.
Anyways, I do use the IPC as a reference to find out about parts and assembly, but since it's neither approved nor accepted, then technically speaking, it can't be used to determine part interchangeability. If we were asked questions on why we installed a part, and we only said it was listed in the IPC, then we could get nailed. I seriously doubt anyone would make a stink about it though, since, as you said, it's a layout of the parts used in a type design. I have no freaking idea why the IPC isn't considered at least acceptable data. Maybe the reason is it doesn't have any oversight or quality control?