Re:The Go Language was rid of blacklist/whitelist and master...
by fanf2 on Sunday June 07, 2020 @09:39AM
(#23447248)
It isn’t clear to me that master/slave has a specific meaning. For instance in the DNS world, it's better to talk about read-replicas, public authoritative servers, update servers, zone transfer servers, DNSSEC signers, etc. - the nuances aren't captured by a primary/secondary distinction. The old master/slave terminology is badly misleading since the DNS protocol doesn't allow the
master to command the slaves to do anything: in the DNS it's the slaves that make the demands. (BIND now supports primary/secondary in its configuration though there are parts that still require the old words.)
Re:The Go Language was rid of blacklist/whitelist and master...
by mcphage on Sunday June 07, 2020 @02:23PM
(#23449399)
> it is also true that it causes a bit of confusion for more experienced engineers.
Experienced engineers are used to handling new and changing terminology. It’s the nature of the job.
I agree that "leader/follower" might be easier to grasp for someone who has never encountered these terms before, but it is also true that it causes a bit of confusion for more experienced engineers.
For example: if I am reading documentation about a hardware controller and find the "leader/follower" terms I would wonder (and, in fact, this has happened to me in the past) if they have the same meaning as "master/slave" or if the author has decided to use these other terms because there is some particular behavior quirk that applies to this hardware that I should be aware of.
All in all, you have to compare cons and pros and, for this particular context (engineering), I think correctness and lack of ambiguity is more important than "feelings"... but this is (obviously) just my personal opinion.