The best engineering practice while making new technologies/procedures is that they look/read into all world current, past, and future related-knowledge
[1] [2], all information phases are important for producing reasonable; I-Ds, RFCs, standards, BCP, etc.
In IETF meetings and discussions there is not way we can ignore the past discussions [3][4]. The IETF is documenting/recording meetings (not reporting attendance size [5]), I-D versions related to RFCs, discussions on the IETF lists, but I think that each WG in IETF after a period of time (about 20 years) can work out an I-D of issues related to RFCs that will be read by IETF next generation engineers (hopefully written by WG RFCs editors before we loose them).
From my experience in reading IETF WGs lists, some authors and WG discussions don't go into reasons of update version of I-Ds, but the reason is only known if a participant asked for it. Some I-Ds authors discuss the change outside the IETF, that is why I suiggested [6].
Please note that many I-Ds and RFC are not perfect described, understood or interpreted [7], because while discussing an I-D in meetings or on lists the interpretation changes by IETF participants (not all readers of that I-D). Therefore, I think there is a need for more historic RFCs in IETF as suggested before [8], but in both procedural and technical IETF works/efforts. I am happy to find [9] in IETF documenting processes (need to be updated) and hoping to see more past engineering helpful experience in I-Ds.
My opinion to make the IETF life easier and linked :)
AB