Add abort to word list

[celestehorgan]

This PR adds the word "Abort" to the word list, as reviewed by the Inclusive Language Initiative's Language Workstream.

Review will be open on this PR until August 31, 2021.

Signed-off-by: Celeste Horgan celeste@cncf.io

[richsalz]

I am not sure about this because abort is a common C function and it's hard for it to not be used at times. I know the recommendation is replace when possible but still it makes me feel a little uneasy.

[Nytelife26]

I am not sure about this because abort is a common C function

We understand that these things can be deep-rooted in places and history, but that doesn't mean we shouldn't try to avoid them in future.

[joannalee333]

Sorry for commenting so late in the process, but I'm not sure why the term "abort" is harmful or exclusionary. I don't believe that use of the term "abort" harms or excludes any historically underrepresented group of people, and when used as a computing term, we're actually using it in accordance with its dictionary definition. I understand that it's a term that is associated with termination of a pregnancy, but I don't believe that necessarily makes the term harmful. This is entirely different from use of the terms "master" and "slave" for several reasons. Everyone agrees that slavery is a horrific and tragic part of human history, whereas opinions differ on whether termination of a pregnancy is an inherently "bad" thing that we should avoid referring to. Similarly, the use of "abort" in computing is a proper use consistent with its dictionary definition of "bring to a premature end because of a problem or fault." In contrast, use of "master" and "slave" in computing does not conform strictly to the dictionary definitions of "master" and "slave" and is unnecessary, unclear, and metaphorical. So I don't personally support deprecation of the term "abort" in computing, but I am open to changing my mind if someone can provide a clear and persuasive explanation as to why the term is exclusionary and harmful.

[Nytelife26]

I don't believe that use of the term "abort" harms or excludes any historically underrepresented group of people

It doesn't have to. It's a loaded term, which is in line with the INI's concerns document.

I understand that it's a term that is associated with termination of a pregnancy, but I don't believe that necessarily makes the term harmful.

It isn't just associated with it, that is actually where the word originated entirely. Its meaning has been diluted into acceptance, rather than diluted into harm. Why have the ambiguity and undertones that come with it and present the distraction to the reader of these documents when there are other better words? There is more rational available in the document, but I hope that helps you understand our thinking :)

And, that isn't just us either. I have since spoken to other women, particularly ones that have had experience with actual abortions, and they agreed it was distracting and not a word they'd use. Ultimately, it isn't something we need to use, and it's loaded terminology - reason enough to replace it.

[joannalee333]

I don't believe that use of the term "abort" harms or excludes any historically underrepresented group of people

It doesn't have to. It's a loaded term, which is in line with the INI's concerns document.

Thanks, @Nytelife26 . I understand that it may be a loaded term, but INI's mission is to "promote and facilitate replacing harmful and exclusionary language in information technology", not to language that is merely socially charged (but is not exclusionary or harmful). So there seems to be somewhat of a disconnect between the scope of the mission statement and the scope of the concerns document.

I understand that it's a term that is associated with termination of a pregnancy, but I don't believe that necessarily makes the term harmful.

It isn't just associated with it, that is actually where the word originated entirely. Its meaning has been diluted into acceptance, rather than diluted into harm. Why have the ambiguity and undertones that come with it and present the distraction to the reader of these documents when there are other better words? There is more rational available in the document, but I hope that helps you understand our thinking :)

I did read the entire document, so I see the rationale. Although I completely respect peoples' decision to choose terms other than "abort" (and maybe there are better terms in many contexts), I don't see it as INI's mission to recommend deprecation of all socially charged terminology or all distracting terminology. I believe the mission is to replace harmful and exclusionary terminology. If a majority of the community wants to issue this guidance to avoid use of the term "abort", I'm not going to oppose it, but just wanted to state my concerns for the record.

[richsalz]

I share @joannalee333's views. I think we dilute our efforts if we're trying to be the overall "language police."

I also dispute @Nytelife26's claim that it came from the term abortion. It came from the IBM term in use for a long time ABEND, abnormal end. But that's not relevant, it's me showing off knowledge of Unix history :)

[richsalz]

FYI, I asked and "abort" in Unix did not come from abortion. At the time, it was in common use (and still is) for things like "abort the mission," "abort the test (flight)" etc. So maybe more military background.

I do not think we should merge this PR.

[joannalee333]

I share @joannalee333's views. I think we dilute our efforts if we're trying to be the overall "language police."

The inclusive language movement is already deemed controversial because some people complain that we are trying to be "language police" and that inclusive language is focused more on political correctness than achieving real change. If we stick to the narrower goal of trying to replace language that is actually exclusionary or harmful and can negatively impact historically underrepresented groups of people and their allies (e.g., master/slave, blacklist/whitelist), I think our effort has more credibility and more of the industry will adopt our recommendations. If we start recommending that all socially charged or loaded terms be replaced, even when we can't articulate a clear case for why the language is exclusionary or harmful, then we give the resistance more ammunition to attack us as being merely about political correctness rather than focused on the more meaningful goals of reducing harm and making tech more inclusive. Also, once we start recommending replacements for socially charged terms that aren't actually harmful or exclusionary, we step onto a very steep slippery slope. If we're going to advocate replacing "abort" merely because it's socially loaded and people find it "distracting", where would we stop?

[markcmiller86]

FYI, I asked and "abort" in Unix did not come from abortion. At the time, it was in common use (and still is) for things like "abort the mission," "abort the test (flight)" etc. So maybe more military background.

I think if we look diligently (I've been trying) at where Unix got its use of abort from, we'd find it comes from ~50's era rocketry research having to do with aborting a launch, which in all likelihood (again, I've really been trying to track down this story) got it from ~1920's era aviation regarding aborting a takeoff or landing. If you look at Google's word usage frequency there is some corrleation with what I am thinking here.

Long story short, someone, somewhere between the introduction of the word (which we all agree appears to have its history in describing the end of a pregnancy by any one of a number of ways) decided to use it for situations more generally, not just the end of a pregnancy, and others followed to mean more generally the end of some process of some kind.

I dunno about you but learning that history impacted where I landed on having INI recommend it be replaced.

Some other things that have come to mind for me regarding this term...

• Does lack of general public knowledge and awareness of that history (I think the origin story here is more obscure than things like master/slave meaning its far less likely very many people actually know this history) suffice to say the word abort doesn't rise to the level requiring INI attention? I don't think lack of knowledge of history is a good argument for that.
• Has the ~500 years of alternative uses of abort since its introduction essentially had the effect of changing the word's meaning from its original meaning? This latter question is, for me anyways, a little more compelling. Multiple current dictionaries provide multiple definitions for this word. One of those definitions is the end of pregnancy. Others are related to ending a process (landing a plane, launching a rocket, running an executable on a computer, etc.). This would suggest that yes, this word has reached a point in usage in common language and those usages have effectively given it new definitions. So, its definition then depends on context. If one is talking about human reprorduction, pregnancy, etc. the use of the word there refers to its original meaning. If the topic is space exploration, its use refers to terminating a launch or a landing or a spacewalk, (essentially some planned activity that is in progress).
• Aborting a process is kinda sorta fundamental to almost all software execution paradigms. Its use is prolific and highly embedded in our software. Yes, master/slave too is prolific but abort is maybe 100x-1000x the usage that master/slave has. So, an INI recommendation here represents a sort of maximal stress test of our activities, IMHO.

I find myself very much influenced by @joannalee333 comments, above. At the absolute minimum, I think @joannalee333 arguments suggest that if INI chooses to tackle abort, it probably isn't the best choice for it to be one of the first items we tackle.

[richsalz]

I think if we look diligently (I've been trying) at where Unix got its use of abort from,

In case it wasn't clear, I asked on the TUHS mailing list (https://minnie.tuhs.org/mailman/listinfo/tuhs) and got responses from a couple of people from Unix v7 days, including Doug McIlroy which I synopsized above. You can stop looking :)

I still strongly believe this dilutes our work.

[markcmiller86]

In case it wasn't clear, I asked on the TUHS mailing list (https://minnie.tuhs.org/mailman/listinfo/tuhs)

Yeah, that wasn't clear. Further, while I myself have no direct contacts regarding unix history, on-line manual artifacts show that the abort() function (section 3, (C) Library functions - November 3, 1971) in Unix-V5 is not present (in fact AFAICT the word abort does not appear anywhere in the V5 manual not even as an nroff command) whereas it is present in Unix-V6 (section 3, (C) Library functions - May 1975) and does appear in other parts of the manual. So, somewhere in that time frame, the practice of using the word to describe abnormal termination as well as the actual C Library function in Unx with that name came into being.

[markcmiller86]

Am putting this note here because it just occurred to me and is highly related but is probably more of an issue for the language workstream to consider.

We talk about harm language causes as one of the guiding principles in INI.

We should probably also have some understanding and characterization of the harm caused by both false-positive (INI recommends replacement of a word that it turns out probably didn't require it) and false-negative (INI neglects to recommend replacement of a word that did require it) decisions.

Managing false-positive and false-negative rates, and establishing (e.g. tuning) a reasonable set point is used in describing receiver operating chracteristics and could be a useful conceptual model in helping INI tune its policies and recommendations.

[RichiH]

No matter what the consensus turns out to be, I think we should list abort alongside the historic context in this issue. Building up a knowledge base is A Good Thing in and as of itself.

[Nytelife26]

I also dispute @Nytelife26's claim that it came from the term abortion. It came from the IBM term in use for a long time ABEND, abnormal end. But that's not relevant, it's me showing off knowledge of Unix history :)

The term "abort" as in "abort mission" was derived alongside abortion specifically to refer to the termination of a pregnancy. You can read more about the etymology of "abort" if you so wish. UNIX history has nothing to do with it, really - the term was made more acceptable by dilution, not made harmful by association, and keeping such a loaded and etymologically diluted term around is a bad decision.

[Nytelife26]

Managing false-positive and false-negative rates, and establishing (e.g. tuning) a reasonable set point is used in describing receiver operating chracteristics and could be a useful conceptual model in helping INI tune its policies and recommendations.

This is a great idea, and one that reminds me a lot of the proselint design philosophy.

[richsalz]

Thanks for the history link, @Nytelife26