On Tue, Jul 14, 2020 at 09:29:39AM -0400, Andy W wrote:
Maybe getting away from the fundamental idea of directionality isn't at
all possible,
well, of course not.
or at least implicit directionality?
huh?
the idea is to avoid suggesting a _particular_ direction (too strongly).
So, with that in mind, how about: "head" and "heel" ?
that's like north and south, only doubly so.
On Tue, Jul 14, 2020 at 01:52:33PM +0000, ARaspiK via isync-devel wrote:
this/that?
no shit - i had that idea as well. then i laughed it off. :-D
here/there?
like local/remote, only more silly-sounding. but see below.
On Tue, Jul 14, 2020 at 07:03:11AM -0600, t...@tkware.info wrote:
How about left/right?
that may sound silly, but this fits the idea of equality *too* well. :-D
(there is slight asymmetry between the sides relating to MaxMessages and
SyncState, and *usually* most data flows in one direction, so i think it
makes sense if the names give a slight hint, like the parent/child
proposal.)
but then, there is this entry in the TODO file:
> - use master:slave for the pattern
i.e., the keywords themselves would disappear from the config file
anyway. however, references in various other contexts would remain, and
left/right would follow the syntax. but so would first/second.
On Tue, Jul 14, 2020 at 09:48:41AM -0400, Stefan Monnier wrote:
Aren't they individually and explicitly given a meaningful name already
anyway?
Endpoint :work-remote:
Endpoint :work-local:
then declaration order would have to define the assignment of the side,
as the relative location is important (pull vs. push). and there would
be also the question how to refer to these in various messages and
command line arguments, as explicitly naming them is not always
practical. so this is actually quite similar to the previous paragraph.
hmm, push/pull actually assume a near/far (new proposal!) mental model,
which is just mildly different from the local/remote one. without that,
the assignment is arbitrary (which it currently is; i remember how i
struggled to get it straight in my head). so, push/pull is up for debate
as well, and proposals which consistently integrate both location and
direction are preferred.
On Tue, Jul 14, 2020 at 04:58:32PM +0300, Roman Bolshakov wrote:
Another alternative for the replica is "peer". No hierarchy implied.
just "peer" doesn't work, because we need two of them.
Some of the changes look like virtue signalling to me because they have
no impact on daily life of the oppressed group while being actively
promoted as such.
you're misunderstanding the term. you're referring to *bad* virtue
signaling. see
https://freethoughtblogs.com/atrivialknot/2016/09/16/virtue-signalling-is-not-pretending/
The renamings don't solve a problem and don't help black people to
suddenly join the industry, rather they distract, divide and confuse
the development community (including developers outside the US).
this isn't about fixing inequality, but about not standing in the way of
fixing inequality. when you insist on using a metaphor that refers to
something fundamentally inhumane, you show that you have no respect for
the people who are rightfully triggered by it, and that certainly
doesn't help. https://cdm.link/2020/06/lets-dump-master-slave-terms/ is
the best article on the matter i could quickly find.
fwiw, my personal boundary are words which are objectively *not* related
to racism, etc., say blacklist/whitelist. however, even in such cases
there are often technical reasons of accuracy and translatability to
prefer something else over a metaphor.
_______________________________________________
isync-devel mailing list
isync-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/isync-devel
