Update to Git changes the default of init.defaultBranch to main

Post by whynaut » Thu, Nov 19, 2020, 2:21 PM UTC

denysvitali wrote: post #139

aaomidi wrote: post #127

Earlier quoted context omitted.

We already assign pronouns and gender to connectors. Seems like you've not actually worked in IT if you don't know that? Pure algorithms have the capacity to change the world. Sometimes for the worse. An example of a pure algorithm with massive political and ethical considerations is crypto currency algorithms. I think a lot of people here would've benefited taking a basic "ethics in CS" class.

I wasn't talking about calling a female connector or a male connector, I was referring to the somewhat new trend of having to specify a pronoun in your Twitter bio because genders are non-binary. > Pure algorithms have the capacity to change the world. Sometimes for the worse. An example of a pure algorithm with massive political and ethical considerations is crypto currency algorithms. Nobody was arguing that. I jus…

Post by whynaut » Thu, Nov 19, 2020, 2:25 PM UTC

da_big_ghey wrote: post #87

mason55 wrote: post #75

Earlier quoted context omitted.

So, two things (and I realize you aren't the person I originally replied to). 1. Your argument here is implicitly agreeing with my point number one. There are terms that would be considered too offensive and so the conversation is about where the line falls. We can't just say "let's not ever talk about whether a term is offensive" because there are some that we really wouldn't want to use, and so we should be talking…

My point is that master/slave actually conveys a relationship between two things. It was picked for technical use because of that. Naming something "nigger", even in a technical would do what, offend some other piece of code? If somebody chose an offensive name arbitrarily (like, say, "git"), I would understand the impulse. Though the former is fraught with specific and awful connotations towards a very specific grou…

Post by dspillett » Thu, Nov 19, 2020, 4:40 PM UTC

jameshilliard wrote: post #151

dspillett wrote: post #84

Earlier quoted context omitted.

> "blacklist" That has already happened in some places, though blacklist/whitelist are still the most common naming convention for that sort of thing. Blocklist is actually a better name IMO, and I've seen it used in places, and an easy enough change. I don't have any examples I remember of what people might use instead of whitelist though, allowlist doesn't particularly trip of the tongue IMO. Grandfathering is anot…

> Blocklist is actually a better name IMO, and I've seen it used in places, and an easy enough change. You should try to avoid using "Blocklist" in most code IMO, it's ambiguous in a number of contexts as the word block itself is ambiguous and can mean very multiple different things such as blockchain blocks or data/filesystem blocks which would be something very different from the intended definition of a disallow r…

Post by sharken » Thu, Nov 19, 2020, 4:42 PM UTC

wayneftw wrote: post #98

tachyonbeam wrote: post #15

Earlier quoted context omitted.

I fully agree that this specific change is insignificant, but the biggest deep learning conference has been renamed, after something like 20 years, because of this kind of pressure. Word policing has a bit of a 1984 feel to it. Some desire to control you and how you think by controlling which words you use, with new arbitrary words being banned all the time. Instead of actual change, we're focusing on the superficial…

> ...the biggest deep learning conference has been renamed, after something like 20 years, because of this kind of pressure. This was interesting: > Of the survey respondents, 1,881 were men and 294 were women. Of the women, 44% agreed a name change should occur and 40% disagreed. Of the men, 28% agreed a name change should occur and 55% disagreed.[5] More than half of those surveyed were against the change, but 30%…

Post by mmis1000 » Thu, Nov 19, 2020, 5:27 PM UTC

WorldMaker wrote: post #246

kruczek wrote: post #235

Earlier quoted context omitted.

> a change as small as a naming convention This "small" change had to be split into 29 parts, so that each part stays below 100kB. This means nearly 3MB of changed code. How many hours did developers have to spend to consider what needs to be changed, make the change and then validate it? In my opinion these are hours wasted, that could be better spent on some real problems.

The post says it was a single global search/replace (probably a grep) operation, plus some automated merge checks. Because most of what was most affected was literally test scripts, almost all of it sounds like it was automated and the vast majority of the "hours wasted" weren't person hours but machine hours.