Stewart: The word master even by itself is still a metaphore for a type of roles by humans. Hence it is exposed to negative connotations that apply to its use in humans. Thats why i was suggesting our rule could be: I) How can we make new work product (RFC) use words that are not gratuitous metaphores without so significanty improving readability as to outweigh possible negative connotations of the metaphores ? E.g.: forget existing use for a moment, and try to figure out what the best word in new work could be. Difficult of course, because at the time of putting the new work together, metaphores from past work will most likely most easily be comprehended, but not necessarily so when thinking about readers 10 yers in the future. Funnily enough, in my observation its only engineering work product where metaphores are so dominant to ease understanding. In patents and research papers, it seems a lot more common to create new words for concepts just to make the work product look more novel. I was always cynical about this "re-branding" (*) but maybe in the context of this discussion it actually helps. And i think readability is core to escape as much as possible the euphemism treadmill problem. Cheers Toerless (*) I have already sent a letter of self incrimination to PETA for the use of the word branding. On Fri, Jul 24, 2020 at 04:12:34PM +0100, Stewart Bryant wrote: > > > > On 24 Jul 2020, at 09:16, Stephane Bortzmeyer <bortzmeyer@xxxxxx> wrote: > > > > On Fri, Jul 24, 2020 at 02:34:18PM +1200, > > Brian E Carpenter <brian.e.carpenter@xxxxxxxxx> wrote > > a message of 10 lines which said: > > > >> Since you ask, the answer seems obvious: its name is "Master Clock" > >> so there's nothing else you can call it. "Master" on its own is here > >> to stay, anyway; in some contexts the proposed alternatives (like > >> "main") simply don't work. That's why most of the advice to authors > >> cannot be binary; we can't resolve this just with a blocklist > >> approach. > > > > And this is also why it cannot be implemented in tools. Even the best > > AI cannot know if the use of a word like master is oppressive or not. > > ... and of course why it cannot be properly litigated against in an RFC. > > The OED defines it as > ???a. gen. A person (predominantly, a man) having authority, direction or control over the action of another or others; a director, leader, chief, commander; a ruler, governor. Obsolete(archaic in later use). > > And when we use it technically we are using it in the same sense but applied between automatons that feel no slight. > > The word goes back to old English but the first quotes use in the OED in the modern spellings is in the King James Bible (1611), so I imagine that it unlikely to be expunged from common English any time soon. > > The first use in conjunction with slavery seems to be 1833 so it is a much more recent definition. > > If you look overall with the OED entry the context in which the authors of the draft raise the term is relatively infrequent, with most associations seemingly relatively benign. > > FWIW the first technical reference to master-slave seems to be in 1952 and in reference to the a system for the manipulation of irradiated items rather than computers. > > The first technical introduction of master-slave I had was the context of a J-K flip-flop where it describes the relationship extremely well. The master takes arbitrary action and the slave has no choice but to follow or very occasionally rebel (metastability is is a big problem in flip-flops) so in many ways the term is a perfect fit. What is the new standard term in that context? > > - Stewart > > > > > > > -- --- tte@xxxxxxxxx