One starts to wonder, at what point might it be actually feasible to do it the other way around, by whitelisting IP ranges. I could see this happening as a community effort, similar to adblocker list curation etc.
replies(9):
I would argue, without any evidence, that when terms are used and embraced, they lose their negative connotations. Because in the end, you want to fight the negativity they represent, not the term itself.
In reply to your argument, the deny list (the actual list, apart from what term we use for it) is necessarily something negatively laden, since the items denied are denied due to the real risks/costs they otherwise impose. So using and embracing the less direct phrase 'black' rather than 'deny' in this case seems unlikely to reduce negative connotations from the phrase 'black'.
It really isn’t. It’s a novel term, which implies a functional difference from the common term. Like, I can run around insisting on calling soup food drink because it’s technically more descriptive, that doesn’t mean I’m communicating better.
To the extent we have a bug in our language, it’s probably in describing dark brown skin tones as black. Not a problem with the word black per se. (But again, not a problem really meriting a linguistic overhaul.)
What do the lists do? They allow or deny access, right? Seems allow/deny are fitting descriptive terms for them then. White/black are much more ambiguous prefix terms and and also come with much more semantic baggage. All in all an easy, clarifying change.
In part. A whitelisted party is always allowed access. If you are whitelisted to enter my home, you always have access. This is different from conditionally having access, or having access for a pre-set period of time.
Same for a blacklist. An IP on a blacklist clearly communicates that it should not be casually overridden in a way a ‘deny-access list’ does not.
> White/black are much more ambiguous prefix terms and and also come with much more semantic baggage
That baggage includes the broadly-understood meaning of the word. When someone says to whitelist an IP address, it’s unambiguous. If someone says to add an IP address to an allow access list, that’s longer and less clear. Inventing a personal language can be an effective way to think through a problem. But it isn’t a way to communicate.
Black and white are colours. (Practically.) I am sympathetic to where folks arguing for this come from. But we aren’t going to solve racism by literally removing black and white from our language.
Arguing that allow/deny or allow/block is less descriptive is basically an argument of "I want things to stay the same because I'm old" or "I like to use jargon because it makes me look smarter and makes sure newbies have a harder time" (and those are the BEST two reasons of all other possibilities)
for those reasons, it's expected that using "black" instead of "deny" will have more support as programmers age and become more reactionary on average, but it doesn't make it any less stupid and racially insensitive
It’s everyone I need to communicate this to already understands what those terms mean.
Also, white and blacklisting isn’t technical jargon. It’s used across industries, by people day to day and in common media. Allow/deny listing would be jargon, because nobody outside a small circle uses it and thus unambiguously understands what it means.
For the same reason, "allow-list" list is not jargon, just like "component" or "extension"
To me there is one issue only: two syllables vs one (not a problem with block vs black for example but a problem with allow vs white) and that's about it.
Of course it is. If I tell someone to allow list a group of people for an event, that requires further explanation. It’s not self explanatory because it’s non-standard.
> just like "component" or "extension"
If you use them the way they are commonly used, yes. If you repurpose them into a neologism, no. (Most non-acronym jargon involves repurposing common words for a specific context. Glass cockpit. Repo. Server.)
Server, cockpit those are jargon. Allow and deny just aren't. Whatever.
Irrelevant since the terms allowlist/denylist do not presuppose conditionallity or pre-set time limits.
> If someone says to add an IP address to an allow access list, that’s longer
Allowlist/denylist (9 + 8 chars) is shorter than whitelist/blacklist (9 + 9 chars).
> Inventing a personal language
Sounds like you think the proposal was to invent a whole new language (or one per person)? I would be against that too. But it is really only about updating a technical industry term pair to a more descriptive and less semantically loaded pair. Win-win.
> we aren’t going to solve racism by literally removing black and white from our language.
Changing to allowlist/denylist would not remove the terms black/white from language. There is good reason for making the change that do not involve any claim that doing so would solve racism.
They don't pre-suppose anything. They're neologisms. So you have to provide the context when you use them versus being able to leverage what the other person already knows.
> Allowlist/denylist (9 + 8 chars) is shorter than whitelist/blacklist (9 + 9 chars)
The point is you can't just say allow list this block of IPs and walk away in the way saying whitelist these works.
> really only about updating a technical industry term pair to a more descriptive and less semantically loaded pair
Eh, it looks more like creating jargon to signal group membership.
> There is good reason for making the change that do not involve any claim that doing so would solve racism
I guess I'm not seeing it. Black = bad and white = good are deep cultural priors across the world.
Trying to bend a global language like English to accomodate the fact that we've turned those words into racial designations strikes me as silly. (The term blacklist predates [1] the term black as a racial designator, at least in English, I believe by around 100 years [2]. If we want to go pedantic in the opposite direction, no human actually has black or white skin in natural light.)
(For what it’s worth, I’ve genuinely enjoyed this discussion.)
[1] https://en.wikipedia.org/wiki/Blacklisting#Origins_of_the_te...
[2] https://nabado.co.ke/2025/01/05/the-origins-and-evolution-of...
Oh I think they do presuppose a link to the main everyday meaning of the terms allow and deny. To their merit! But yes they do not presuppose conditionality or time-limits.
> versus being able to leverage what the other person already knows
I'd guess over a million people start learning software dev every year without any prior knowledge of these industry terms. In addition while dev terms often have english roots many, maybe even a majority, of new devs are not native english speakers, and for them the other meanings and etymology of whitelist/blacklist might be less familiar and maybe even confusing. In that regard allowlist/denylist have a descriptive advantage, since the main everyday meaning of allow/deny are mnemonic towards their precise technical meaning and when learning lots of new terms every little mnemonic helps to not get overwhelmed.
> you can't just say allow list this block of IPs and walk away in the way saying whitelist these works.
You can once the term is adopted in a context, like a dev team's style guide. More generally there can be a transition period for any industry terminology change to permeate, but after that there'd be no difference in the number of people who already know the exact industry term meaning vs the number who don't. Allowlist/denylist can be used as drop in replacement nouns and verbs. Thereafter the benefit of saving one character per written use of 'denylist' would accumulate forever, as a bonus. I don't know about you but I'm quite used to technical terms regularly getting updated or replaced in software dev and other technical work so this additional proposed change feels like just one more at a tiny transition cost.
> it looks more like creating jargon to signal group membership
I don't think any argument I've given have that as a premise. Cite me if you think otherwise.
> The term blacklist predates
Yep, but I think gains in descriptiveness and avoiding loaded language has higher priority than etymological preservation, in general and in this case.
> Trying to bend a global language like English
You make the proposed industry term pair change sound earthshaking and iconoclastic. To me it is just a small improvement.
Thanks for the discussion!