From my Jewish perspective, I think of it as oral tradition vs written tradition (like the pre-Mishnaic distinction between the Written Law and the Oral Law).
In general, I see good commit messages as a Talmudic codification of a company's oral knowledge - complete with back-and-forth conversations, contradictory points of view and distinct characters. Still better than having to go talk to the people involved, some of whom may no longer be in the organization.
(Specifically with commit messages as Gemara and wikis as the much more organized and impersonal Mishnah.)
It is still surprise to me the switch to woman based Jewish identity and the importance of oral tradition in your culture. Also shocked when Dead Sea Scroll come out, I only knew then the Jewish bible oldest is only 1000 years old.
There’s a theory that all written artifacts are at most 1000 years old. Christian dates (in Roman) got M (1000) infront of all dates and Jewish an V (5000).
So we got 2000 years and 6000 claimed history.
Having written the above it appeared to me this could be an ironic illustration of tribal knowledge vs technical documentation issue. Knowing when something happened and reading about it are two separate things.
I'm sure this is a valid point, but I think it is overemphasized.
Much of that tribal knowledge would have been very esoteric and focused on technologies that are no longer critical. for example, fabricating magnetic core memory. I'm sure some of the required skill is gone, but we don't need to use that memory technology anymore.
having been involved in spacecraft operations, it has nothing to do with how the memory is wired up.
there would've been a huge amount of knowledge about how to operate apollo spacecraft in general. there is no way it is preserved.
and this thread is beneath a comment suggesting that we don't start from scratch, and reuse apollo designs. as you point out, we can make better stuff now. the only value to the apollo designs would be if we had operational experience to go with them.
Tell that to Boeing. Thier last flight didn't make it to the space station because of two computer errors. Had that happened over the moon, people likely would have died.
Hm, it seems I should have been more specific. I am aware of the mechanism, I'm hoping someone might have even a bit of insight into what that tribal knowledge is/was.
> Finally, the parachutes were folded and packed by hand. During the Apollo missions in the 1960s and early 1970s, only three people in the country were trained, and then licensed by the Federal Aviation Administration, to fold Apollo parachutes—Norma Cretal, Buzz Corey and Jimmy Calunga —and they handled all 11 Apollo missions. Their skills were considered so essential that NASA forbade them from ever riding in the same car together. The agency couldn’t afford to chance that all three would be injured in a single accident.
I also doubt that the seamstresses techniques to sew those parachutes are something easily written down
I don’t think that’s the right way to phrase it. Complicated things are probably written down somewhere, but it’s just not practical to assimilate the knowledge without having been involved in its creation.
The loss of this kind of thinking isn’t so much a failure of people to write documentation, but a natural consequence of things being hard.
> Isn’t it odd that the most important political and military mission of a generation relied on “tribal information which wasn’t written down?”
No. Figuring out how to accurately describe something in writing is hard, and then there are all the details you have to leave out to make the document readable, some of which may be more important than you realized.
Every project relies on tribal knowledge, and much of that knowledge will be lost unless is followed up by an equally massive project by a horde of technical writers to nail everything down.
They have to be readable enough for the reader to be able to recreate what's described from those documents, and you can only test for that by... trying to use the documents to recreate the work. Which is rarely attempted, especially in large, expensive projects.
(It's the same thing as the "you only have a backup if you actually restored it at least once" adage.)
Especially in an era where you had to literally write things down versus throwing in a link to a slack message onto a Jira ticket somewhere (Which no one will ever read anyway since text search in Jira sucks)
Also its worth emphasizing that a lot of the actual physics regarding like rocket combustion was not understood so the design process had elements of "make a rocket engine that satisfies some generic constraints and we can figure out the black magic parts such as POGO". Not that useful to write down.
