Request that the keywords "husband" and "wife" be protected from deletion

@keyword_expert​ 

Wouldn't you say that a title with the keyword husband-wife-relationship still needs husband and wife as keywords? As with cigarette and cigarette-smoking, there's a lot of overlap, but not complete overlap.

A husband can appear in a title without his wife, and a wife can appear without her husband. And then we also have husband-husband-relationship and wife-wife-relationship. A title with one of these keywords may also have only a husband or only a wife in it.

A title with the keyword mother-son-relationship needs also to have mother and son as keywords. Especially mother. Someone searching for titles with any kind of mother in it may not care whether she's the mother of a son or a daughter. They may just be looking for mothers.

I'm against eliminating keywords that overlap with other keywords, unless the overlap is complete. Even if there's a 99% overlap, I still like to see both keywords in existence and placed on all titles where they are appropriate.

I'll take a screenshot of the titles where the keywords were deleted and reply again.

@jay_spirit​ You make some very good points. I did think about these exact questions before joining in the mass editing of "husband" into "husband-wife-relationship." I figured this would not be a controversial edit and that the benefits of having the more specific (and more prevalent) keyword would outweigh any negatives. But you are correct there is not 100% overlap between these keywords.

I myself previously added "husband," "wife," and "husband-wife-relationship" all to a particular title (which is why I noticed when that title was recently edited). Then once I observed the mass editing, I joined in on a limited basis. 

Rest assured that you have convinced me not to delete or edit "husband" or similar keywords from titles in the future.

The harder part will be convincing our other fellow contributor to cease these edits. Sometimes even when this person agrees to change certain keyword editing practices, that same pattern later resumes. 

@keyword_expert​ 

Here's a screenshot of the titles with inappropriately deleted keywords. I screenshot the submissions where I restored them. As you can see, in one case, a keyword was deleted yet again.

I am on my phone right now but when I get home I will check my records for whether I edited any of these titles. 

@jay_spirit​  I am home now and checked my records. I only edited 1 of the 11 titles. It was the title Out of the Darkness, on which I changed "husband" to "husband-wife-relationship." I haven't ever deleted or revised any keywords on the other 10 titles. Those other titles must have been edited by the other contributor.

@jay_spirit​ p.s. And I should clarify that my edit to Out of the Darkness was over 48 hours ago. It looks like that title was edited again today based on your info. 

@jay_spirit​ 

A husband can appear in a title without his wife, and a wife can appear without her husband.

BTW, I thought of better keywords than "husband" and "wife" for these specific situations: I think "married-man" and "married-woman" are better. That doesn't make "husband" or "wife" inaccurate or inappropriate keywords, but I do think "married-man" and "married-woman" would be better.

@jay_spirit​

Damn. 

The keyword "husband" is now down to 3 titles.

https://www.imdb.com/search/title/?keywords=husband

When I was "auditing" the keyword this weekend, I got it down to about 100 titles, but was unable to get it lower than that. For many of the titles it was unclear if it should be "husband-wife-relationship" or something else entirely. 

@jay_spirit​ You might want to reiterate your request to IMDb staff to "protect" the keywords "husband" and "wife" so they can't be easily deleted. It is possible to protect certain keywords from deletion or to set them up so they require human review of any deletions. 

@keyword_expert​ 

Thanks. I edited the parent post to make that request.

@jay_spirit​ I wouldn't say it's completely useless. I successfully use the keyword search on an almost daily basis. 

But you're right that some contributors are a bit overzealous with their interpretations of what exactly a "duplicate" keyword is.

I just found out why someone deleted stepson-kills-stepfather. The contributor wants to replace it with stepson-murders-his-stepfather.

Someone who kills another person is not necessarily committing murder. There is killing for self-defense. There is accidental killing.

Stepson-murders-his-stepfather would overlap with stepson-kills-stepfather, but does not replace it.

Yet another foolish "correction."

@jay_spirit​ Yes! I agree with you completely. That confirms for me that all of these edits were done by the same person I was thinking of, because I have observed in the past that this person also fails to respect the differences between killing someone and murdering someone. I have noted this exact keyword editing issue before.

Please keep me posted. 

@jay_spirit​

For titles where a stepson murders his stepfather, it is totally okay to have both the keywords "stepson-murders-his-stepfather" and "stepson-kills-his-stepfather."

That way, people can search (and sort their searches) based on the broader "kills" keyword or the more narrow "murders" keyword, depending on what they want to search for.

The thing that really boggles the mind is this particular contributor also creates thousands of what I call "orphaned" keywords -- keywords that will only ever apply to one title. Here is what I am talking about:

What's up with all the highly specific keywords?

It seems hypocritical to try to reduce excess keywords like what we have observed (e.g., deleting "amnesia" because that concept is part of another keyword), and yet at the same time spamming walls and walls of orphaned keywords that will be functionally useless in multi-keyword combination searches (because they will only ever apply to one title). When submitting a keyword, the "tips for keyword submissions" even advises contributors to try not to create new keywords unless absolutely necessary:

Before submitting, search the keyword database to see what already exists - if it does, please use it instead of submitting a new keyword.

@jay_spirit​ I should clarify that I don't know one way or the other whether the "amnesia" keyword was deleted from that title you mentioned by the same contributor I have in mind.

Often with keyword additions, revisions, and deletions it is possible to determine which contributor is responsible, based on context clues and also because most prolific contributors have their own signatures or indicia in their keyword work. But with the deletion of the keyword "amnesia" on that particular title, I just can't tell.

All the other keyword changes we have been discussing in this thread, however, were almost certainly done by the same top contributor.

@keyword_expert​ 

I restored my submissions yet again.

Allowing cranks to make whimsical changes and deletions to the database is disgraceful. I hope the staff will do the right thing and put an end to this.

I hope the staff weigh in on this soon. 

The contributor is deleting my work for the fifth time, as I write this.

@jay_spirit​

So lame.

Are you adding the keywords manually, or en masse?

Hopefully the latter, because it's a lot quicker that way. 

Adding keyword to multiple titles

@keyword_expert​ 

I'm in favor of it.

I'm surprised the keyword family still exists. I thought all the keywords that duplicated the names of genre categories had already been eliminated.

@jay_spirit​ Yeah, that's exactly what caught my attention about the "family" keyword. And by the way, in my other post I also proposed changes regarding mergers into the "stepfather" and "stepmother" keywords, so I'm really getting a strong feeling that it was my other thread that triggered the mass edits of these keywords (including "stepfather" as shown on one of your screenshots). Often while my proposed mass mergers are pending community discussion and awaiting action by IMDb staff, one contributor in particular engages in manual mass edits of his own, bypassing the community discussion. Nothing I can do about that except point it out on this board.

Years ago, contributors complained that a keyword like killing dog was too ambiguous. Does it mean killing a dog or a dog that kills?

And so everyone agreed that a keyword with that particular gerund/noun combination should have an article in the middle. Not killing dog, but killing a dog.

But one contributor decided with no one's permission that any kind of clipped, headline-style speech must be removed from the keywords.

He made tens of thousands of wholesale changes.

Now, we can't have husband kills wife. It must be husband kills his wife. We can't have hit on head with bottle. It must be hit on the head with a bottle.

It's too late now to go back to the headline style, but I found it easier.

Too bad one contributor can make such a mess.

I have noticed those differences in keyword styles (differences between the keywords with and without articles like "a" in the middle). A lot of the old style keywords do still remain, though. I come across them all the time.

In most cases (but not all cases) I think the newer style makes more sense, and that is usually what I endorse. For example, "sex-in-a-car" is better than "sex-in-car," and "mother-slaps-her-son" is slightly better than "mother-slaps-son." (Plus, as you point out, that is the direction that the keywords have gone.)

With the auto-conversions that are resulting from my posts, users can enter whichever style they prefer ("husband-kills-wife" or "husband-kills-his-wife") and it will be automatically converted for them. The merger and auto-conversion method is the preferable way to go, because going forward, it won't be necessary for contributors to keep making thousands of manual edits.

But the bigger point is that it's better for wholesale changes to result from community discussion. Personally I feel like if someone is willing to spend hours each day tinkering around with keywords by manually adding "a" to the middle of thousands of keywords, then more power to them. But it would be preferable for that to be a community effort.

@keyword_expert​ 

I agree. The tinkering is a waste of everyone's time. Any wholesale corrections or deletions of keywords should raise a red flag to the staff.

In the bad old days, we had no choice. I remember making wholesale changes myself that I later regretted. But now we can appeal to the staff to make them.

The democratic way you do it is much better. It weeds out bad ideas. And everyone has them sometimes.

@jay_spirit​ Definitely. If I have a bad idea, I want to be called out on it. The last thing I want to do is merge a good keyword into a bad keyword, especially in a way that can't be undone. That has happened a handful of times even with my community-vetted mass mergers, and each time was a lesson learned.

I'd like to see the staff flag for review anyone making too many corrections or deletions of keywords.

A contributor who used to post here wanted to remove the keyword horse from every western because he thought it was redundant.

Too many contributors have a basic misunderstanding of what keywords are for. Even some staff members do.

No one should make wholesale changes to keywords without a general discussion.

@keyword_expert​ 

The contributor has stopped making the deletions for now. I'm considering marking this solved, and starting a new thread requesting that IMDb forbid any kind of wholesale changes to keywords.

They might consider making this a rule in their guidelines and flag anyone who is making too many changes to a specific keyword.

What do you think?

@jay_spirit​ I would oppose such a rule, because I do a lot of mass cleanup (aka "auditing") of bad keywords of my own, and I know the other contributor does some good work like that too. Unfortunately I don't think it's possible to write a rule that explains what types of mass keyword editing is good and what types are bad. I definitely wouldn't support a blanket prohibition against wholesale changes. 

I did notice that the contributor has been leaving "husband" alone. Perhaps the contributor finally read this thread.

I wouldn't mark the thread as "solved" just yet, though. I still think it would be good for staff to weigh in.

I decided to boost "knock-at-the-door" by a couple hundred titles, and also saved the first page of 50 titles in the Wayback Machine.

https://web.archive.org/web/20220926051338/https://www.imdb.com/search/keyword/?keywords=knock-at-the-door

https://www.imdb.com/search/keyword/?keywords=knock-at-the-door

@jay_spirit​

I have noticed this too. In most cases I support changing "the" to "a" in keywords. For example, I never understood why "fired-from-the-job" was such a popular keyword, and I was able to get it changed to "fired-from-a-job" after a community discussion.

But there is another important difference between "knock-at-the-door" and "knocking-on-a-door," and that is the difference between the words "knock" and "knocking." The first keyword refers to the knock itself, usually when it is heard from inside the door and the knocking itself cannot be seen. The second keyword is when the knocking is actually shown. 

The other clue is the different words "at" and "on." We often say "who's that knocking at my door?" because the word "at" emphasizes that the knocking is being heard from the inside, and the door presents a barrier from the outside. The first dictionary definition of "at" applies here: "(used to indicate a point or place occupied in space); in, on, or near: to stand at the door; at the bottom of the barrel."

If there is a better keyword than "knock-at-the-door" I am open to that. But there needs to be a differentiation between knocking on the outside of a door and hearing the knocking from inside the door.

I thought about maybe "a-knock-at-the-door" but it seems odd to add "a" to the beginning of a keyword.

Another possibility would be "hearing-a-knock-on-a-door." But that doesn't make much sense to me either, since the focus should be on the knock itself rather than a character hearing it. 

This is another good example of why it makes little sense for one contributor to make unilateral decisions about mass deletions of a specific keyword, unless it's very obvious that the keyword is improper. I hope our fellow contributor is reading this. 

Plus, we live in an era whereby it may be a bit much to assume that a husband is necessarily a "he" or that a wife is necessarily a "she", especially when considering gaps or unknowns in the origins and etymology of words pertaining to social status. For ages, the husband has signified the senior/lead partner in a marriage, whereas the wife or wives signified a person or people belonging intimately under the husband's leadership. As per emancipation, the husband had to be a grown man. As per men's entitlement to be free to a degree, the wives had to be women and girls. The ancient Spartans presumably used two words to distinguish between a man's male (protégé) plaything and his female (procreative) partner, but I couldn't say for sure.

@jeorj_euler​ This is a good point. The very same thought has crossed my mind more than once.

I do wonder if we would be able to identify any title on IMDb where a female character is considered to be a husband, or a male character considered to be a wife. If anyone can identify such a title, perhaps we should reconsider inserting "his" and "her" into all these keywords.

To be honest, I've only ever observed the colloquial term "prison wife" being used euphemistically by male inmates in interviews incorporated into a documentary, the title of which I cannot remember. When they speak of this, they're not even necessarily talking about transgender women, not even men in drag, just a man locked into general servile role that does not exclude sexual behavior.

@bradley_kent​ 

May have, but don't remember deleting "husband" or "wife" as keywords.  I do wonder if they should be deleted and banned entirely.  Can one BE a husband without a wife, or husband, and a wife with a husband, or wife? Don't "husband-wife-relationship," "husband-husband-relationship," and "wife-wife-relationship" cover the same info? 

This is covered extensively in this thread. Among other reasons, "husband" plus "wife" simply does not equal "husband-wife-relationship." As @jay_spirit and you have both pointed out, a marriage can involve two husbands, or two wives, or a husband and a wife. For that reason, "husband" + "wife" and "husband-wife-relationship" are not duplicate sets of keywords. Jay helped me realize that. Someone might want to search for all titles involving husbands, and some people use the keyword "husband" instead of "married-man." (I would be open to all instances of "husband" being changed to "married-man," but only after a community discussion of whether that is appropriate.)

The way keyword searches work on IMDb, if someone wants to search for all titles involving husbands (in combination with other keywords), there is no way to do that with a keyword search other than the keyword "husband." It literally does not work to try to do a search that would use "husband-wife-relationship" OR "husband-husband-relationship" in place of the keyword "husband." There is simply no way to do this.

Whether it was you or someone else who started the trend of mass manual deletions and edits of these keywords, I would encourage whomever it was to simply admit the mistake and/or misinterpretation of the keywords and guidelines, and even more importantly, to work to repair any damage done.

I make mistakes all the time. It can be liberating to admit them.  I did exactly that in this very thread. I had temporarily joined in the mass manual deletions of "husband" and "wife," and I freely admitted that, but eventually Jay helped me realize that I was wrong to do so.

We are all human, and to err is human. Nobody is perfect.

Several months ago,  tried to convert keywords like "18-year-old-teenage-boy" into "18-year-old-boy." Thought I had done a good job, but apparently not. 

Thank you for that cleanup work. Here is a link that will help you find the remaining keywords of that nature. 

https://www.imdb.com/find?s=kw&q=year+old+teenage+&ref_=nv_sr_sm

Most of the keyword problems discussed seem to be attached to (apparently) student films, short films, short and feature Russian films, short Czech films, etc., which suggests that they are coming from the same sources.

I'm not sure which "problems" you are referring to, but if you're referring to the "husband" and "wife" keywords and "kill" vs. "murder" keywords, these are system-wide changes involving hundreds of titles. They are not limited to only one type of title. Over the years, one contributor has been "curating" certain keywords, no doubt believing they were doing good things, and in many cases I believe they were contributing valuable cleanup work. But in some cases this mass manual work was, at a minimum, a waste of time and effort, in some cases it quite literally canceled out the good work of their fellow contributors, and in worst-case scenarios it has actually created inaccuracies in the keywords, some of which still linger.

Whether or not it was you, any help cleaning up these errors would be greatly appreciated.  A good example is the "murder" vs. "kill" keywords.

Right now I am still working to try to clean up the "characters-murdered-one-by-one" keyword. Specifically, the keyword "characters-killed-one-by-one" was changed to "characters-murdered-one-by-one" on dozens of titles over the years. For example, this even happened on Jaws and Jaws 2, even though a shark cannot murder a human. 

Here are some links that can help identify titles where "characters-murdered-one-by-one" (and potentially other "murder" keywords alongside that keyword) have been improperly applied.

https://www.imdb.com/search/title/?keywords=characters-murdered-one-by-one%2Ccreature-feature&ref_=kw_ref_key&sort=moviemeter,asc&mode=detail&page=1

https://www.imdb.com/search/title/?keywords=characters-murdered-one-by-one%2Ccreature&ref_=kw_ref_key&sort=moviemeter,asc&mode=detail&page=1

https://www.imdb.com/search/title/?keywords=characters-murdered-one-by-one%2Calien&ref_=kw_ref_key&sort=moviemeter,asc&mode=detail&page=1

https://www.imdb.com/search/title/?keywords=characters-murdered-one-by-one%2Czombie&ref_=kw_ref_key&sort=moviemeter,asc&mode=detail&page=1

When in doubt, the keyword "characters-killed-one-by-one" is broader, so it can be used without resulting in any inaccuracies. But for titles where "characters-murdered-one-by-one" applies, there is nothing wrong with using both keywords, because when a character is murdered, they are also killed. Using both keywords facilitates multi-keyword-combination searches, which in my opinion is the best and most important function of keywords.  

 Perhaps I am mistaken, and giving the staff too much credit, but I follow the assumption that, if the contribution of an addition, correction or deletion is "accepted," it should usually be considered "acceptable."

Please do not approach keywords with this mindset. The vast, vast majority of keyword deletions and edits are approved automatically, within seconds or minutes, without any human staffer ever seeing them. That fact should not be interpreted as approval by IMDb staff. This is especially true with us Top Contributors. Those of us who have made hundreds of thousands of keyword-related contributions have built up a certain level of trust, where our edits are accepted automatically and staff assume they are correct. This gives us Top Contributors a lot of power. And with great power comes great responsibility. 

@keyword_expert​ 

Thank you very much for this explanation.

I wasn't fully aware of these differences, and also because of missing or fewer keywords in different areas where these "...-killed" variants occur, I didn't want to introduce new ones - e.g. "murder-of-..." or "...-murdered".

In many categories I have already seen imprecise definitions such as "...killed", which are correct for the time being, but could still be made more precise and turned into "murder-of-...".

Over time, however, I will change them or set new ones, as I have noticed that these differences can be important here.

I'll go through your lists over the next few days and can correct or add keywords where I've seen the films myself and know for sure.

There are also films in which maybe both variants, i.e. "murder-of-..." and "...-killed" are correct, since in many action films multiple characters die in different ways.

@DataOrganizer​ 

So if i understood it correctly, an extraterrestrial being cannot commit a "murder". But what if the film takes place in the future or in a parallel world in which extraterrestrials are integrated into society and are therefore also subject to the law? Wouldn't it be murder in this case - or not, since it's a "fiction"?

Yes, that is exactly right. There can be rare cases where an extraterrestrial alien (or a vampire, or an undead character, etc.) can in fact commit a murder. This will require a careful analysis of the plot for each title.

There are many other instances where killings are not murders. A person infected with a virus or possessed by a curse or evil spirit who is not in possession of their own mental faculties may kill someone without murdering them. A killing that happens as part of a war is (usually) not murder. An execution approved by a government is (usually) not murder. A character who kills a dying character out of mercy (a mercy killing) is (usually) not committing murder. And there can always be exceptions to the exceptions. 

Even for titles where "murder-of-..." is appropriate, you can also use the counterpart "killing-of-..." keywords on the same title. That is because all murders are also killings.

But when in doubt, the "killing-of...." keywords are always best, because they are broader.

Thank you for taking a look at these keywords. I have been doing the same. Together we should be able to fix any problems. 

While the lone Xenomorph in Alien may be sapient, it is also feral, so it wouldn't be subject to the same law governing/protecting the Nostromo crew members. So to speak, murder can sort of only be committed by an entity that would ever even be entitled to due process of law as to the legal mitigation of that entity's behavior upon it being captured or tamed; a far cry from being summarily put down like a rabid wolf that has already claimed the life of a human being.

So, as to the matter of vampires (or werewolves for that matter); given the way that most stories concerning their relationship with society are constructed, generally when a vampire claims the life of a human being, or conversely when a human being destroys a vampire, it does not count as murder. The prevalent theme in these fictional universes is that the townsfolk will emerge with their pitchforks and torches to slay the vampire, regardless of how many offenses it/he/she committed, and have impunity in doing so. Absent the fervent belief that the body of the slain was human rather than vampire, no prosecutor's office will charge the townsfolk with murder, and nobody will hold the office in dereliction of duty. Just as well, law enforcement's reaction to a vampire who has slain a human being is typically equivalent with they way "townsfolk" or societies do. If the vampire surrenders, the public authorities will find some way to summarily destroy it/him/her, unless the law (backed up by force) clearly prohibits such. Now, if there is a story whereby clarification is provided that the law does govern/protect vampires in the contexts of homicide, then the murder concept may indeed apply. This would probably make the most sense as a big treaty between human nations/tribes and vampire covens, but certainly it could just as well be that one kind has supremacy over the other kind yet requires that extensive protections be provided to the more "oppressed" of the two kinds. The same principle applies to other sapient creatures or, for example, mecha/software.

The matter could be even more obfuscated or at least interesting in regards to witches, warlocks, wizards, sorcerers and X-men-type of mutants, or superpowered humans (inhumans, metahumans or so), since they are indeed essentially human beings (probably members of society too) and, unlike vampires, do not need to prey upon human bodies/minds in order to survive, thus wouldn't be seen as monsters. Typically these kinds are portrayed as having a better shot at being protected by the law, not to mention that they are more likely to be raised/indoctrinated to follow the law in the first place, even openly serving in public offices too (or ruling society) without a much of secret agenda out of the ordinary.

O. In all this thinking I've been doing about this keyword matter surrounding "killing", i.e. homicide, I completely forgot to mention the concept of manslaughter. A killing, if done by one human being to another human being, could be murder or manslaughter, or even another thing out of some half a dozen concepts with distinctive descriptions of the behavior, intent and motive.