Duplicate Keywords - List #53 (real-life people keywords pt. 2) (Proposals for Permanent Merger and Auto-Conversion)

@Peter_pbn​ 

Thanks for the heads up, and also for the reminder about that list. I see a couple other bad keywords that should really be taken off the list, including "multiple time frames" (it should be spelled "multiple-timeframes"), "no background score" (a poorly formatted keyword that probably shouldn't exist), and "no music" (same), and "cult" (people should not be using that keyword for cult films, but rather for plots involving actual cults). 

I still believe "shakespeare" should be merged into "reference-to-william-shakespeare." There are other keywords like "shakespeare-adaptation" and "based-on-shakespeare" but none of these keywords have been widely used. 

I think I would recommend just removing some of these keywords like "shakespeare" and "no-music" from that list of "Interesting Keywords." It's about time for an overhaul of that list, anyway. I will probably start up a new thread about it.

It is not just about that list, but about the mobile app keyword feature, which has a separate list. Changing the keyword without updating the app would presumably cause an error. None of the other keywords discussed here appear in the app.

In the list of interesting keywords is also multiple storyline, which has previously been changed to multiple-storylines on your suggestion.

I'm not a contributor of no background score or no music, but since most movies have scores or music, I think it is notable when they don't. IMDb editors apparently agreed.

cult being included in a list of keywords doesn't suggest that it should be used about cult films.

PS. I would probably have started a post about it eventually, but you are free to do so.

@Peter_pbn​ 

I am not familiar with the mobile app keyword feature. Can you give me more details?

I removed "shakespeare" from this list for now. That keyword should probably be "audited" first anyway, since most instances of "shakespeare" should probably be something like "shakespeare-adaptation" rather than "reference-to-william-shakespeare." 

I continue to believe that "no-background-score" and "no-music" are bad keywords, but admittedly I haven't looked into it very much yet.

And while it is technically correct that the inclusion of the keyword "cult" doesn't denote any specific intention or outcome on its face, the unfortunate reality is that the keywords "cult" and "cult-film" are included side by side on this list of "Interesting Keywords," and contributors do tend to (incorrectly) use these two keywords together. There are currently 216 titles with both of these keywords together, and many hundreds of titles with the keyword "cult" where no actual cult is involved in the plot.

Plus, combine those points with the fact that the "Interesting Keywords" list we're talking about is heavy on category/style/genre keywords like "farce," "epic," "satire," "anime," and "neo-noir," and relatively light on actual plot keywords like "biker," "robot," and "ninja," and it's very likely that the IMDb staff included "cult" on that list because they were thinking of "cult movies" and "cult TV" (and the keyword "cult-favorite") rather than actual cults (which would have been an erroneous interpretation of what the "cult" keyword should be used for). Of course, I can't prove that definitively. But either way, the "cult" keyword should be removed from the list because it is contributing to the misuse of this keyword. There are plenty of better keywords that could take its place.

The app's advanced search just has a list of less than a hundred keywords to choose from.

I have honestly been reluctant to start posts about keywords while you have been posting so many, but I could start one to ask about shakespeare so that it can be resolved.

@Peter_pbn​ 

I wrote a new post about the "Interesting Keywords" list.

I think it makes sense for you to also write a post about the mobile app, and any keywords (including "shakespeare") that should be removed from that list. It's a separate topic that deserves a separate post. 

@bradley_kent​ 

The guidelines preclude listing people's names by themselves as keywords. Unfortunately, many people are unaware of the guidelines and violate this one all the time. This has been a problem for many years, as shown in this post by @Marco.

The idea behind this list is to look for the more prevalent improper keywords that fall into this category, and nip the problem in the bud by preventing these keywords from being created in the future. The alternative -- perpetual manual editing of these keywords a little bit at a time over many years -- is like a Sisyphean task of pushing a boulder up a hill, only to have it roll back down all over again. This approach also creates a false sense that the problem has been solved, when any solution is usually just temporary (until the bad keyword is re-created). And worst of all, if these keywords are not converted to "reference-to-" keywords, then that reduces the ability to use the "reference-to-" keywords in keyword searches, including keyword combination searches.

As we have discussed before, there is absolutely nothing wrong or "incorrect" with using a "reference-to-" keyword on a title that also has a character by the same name as that keyword, or a documentary about that person, or a title starring that person as herself or himself. This facilitates keyword searches, because it allows people to use the more prevalent "reference-to-" keywords to look for titles and patterns involving the person referenced in that keyword. After all, even a title with a character of a real person (or starring that real person as themself) also references that person. These are examples of appropriate uses of "reference-to-" keywords.  

I would thus describe the use of "reference-to-" keywords on these titles not as incorrect, but rather incomplete. But keyword sections of titles are always incomplete -- there can always be additional keywords added to them. 

On that note, nothing is stopping you from adding "-character" keywords to these titles (where applicable) in addition to the "reference-to-" keywords. It sounds like you are already doing that (or more likely, you are only applying "-character" keywords and not "reference-to-" keywords, which is unfortunate, but that is your choice). 

@bradley_kent​

In Ballot Box Bunny (1951), the use of the Teddy Roosevelt quotation "I speak softly, but I carry a BIG stick!," accompanied by the brief parody where Bugs Bunny is dressed like Roosevelt, also qualifies this title for the keyword "reference-to-theodore-roosevelt."

This is what's called an allusion. The dictionary defines "reference" as "an act or instance of referring" or "a mention; allusion." And "allusion," is defined, in turn, as "a passing or casual reference; an incidental mention of something, either directly or by implication: an allusion to Shakespeare."

Thus, Ballot Box Bunny is referencing Theodore Roosevelt by allusion or implication, which makes the keyword "reference-to-theodore-roosevelt" 100% accurate and appropriate for this title.

Incorrect, standalone proper name keywords are much easier to "zero in on" and correct.  (I have personally corrected thousands of them.). By merging them into "reference-to-" keywords, you are expanding the number of "reference-to-" keywords that would need to be audited, doubling, tripling, quarupliing, etc. such keywords into a greatly expanded list that would probably never be audited and, therefore, not corrected.

@keyword_expert​ This remains a terrible suggestion.  Some of these proper names need the (proper-name)-character keyword.  To convert ALL of them to reference-to- keywords is a grave mistake.

Again, only a careful audit of each standalone proper name keyword will solve this problem.  By looking at the cast list on EACH title, one cam determine IF the (proper-name)-character keyword is needed or IF the reference-to-(proper name) keyword is needed.  By converting ALL to a reference-to- keyword, you are just turning the problem into an even bigger one.  (Look at "Carry on Columbus" for just one of many examples.)

@bradley_kent​ 

At this point I will simply refer you to our original conversation on this topic. 

@keyword_expert​ 

The problem is that thread doesn't actually cover the fact that you cannot assure that all of these are "references" to a person and not the actual character appearing. If you believe that a character occurring in a feature is enough to say it is a reference, that is kind of a twisted logic.

When I see these keywords, a lot of time they are meant as character keywords, not references. today I changed one for "the-falcon" to "the-falcon-character". Changing these all to references wouldn't do much for clarity.

@adrian​ 

The problem is that thread doesn't actually cover the fact that you cannot assure that all of these are "references" to a person and not the actual character appearing.

Yes it does. Thoroughly. I encourage you to read the full thread.

I am in the middle of two other large IMDb projects, but have just taken the time to do a beginning audit of the 52 "christopher-columbus' keywords.  Just going by the cast listings, I have been able to correct 21 to the correct "christopher-columbis-character" keyword.  Of the remaining 31 titles, more research is needed to determine the correct keyword.  By just merging these 52 keywords into the already existing (and in need of auditing) 122 "reference to-christopher-columbus" listings,  you will be compounding the future auditing need.  It is easier to research 52 keywords than it is 174.

This is why the incorrect (and remaining) standalone proper name keywords should remain until they can all be accurately and carefully audited.

P.S. The "christopher-columbus-character" keyword now has 34 correct title listings, with the acknowledgement that there will be future correct additions following research.

@bradley_kent​ 

This list has already been pending for two months, and it will be longer still before IMDb staff will act on this list of duplicate keywords, for a number of reasons (the holidays, the backlog of pending lists, and the fact that staff do not currently have the ability to do mass mergers while some adjustments are made to the site).

Plus, you also have the option of saving all the titles you're interested in in a private list for future review and action (following the mass mergers, whenever they do occur). 

Otherwise, I reiterate that "reference-to-" keywords on these titles is not "incorrect." The question is whether any of the titles could also use "-character" keywords or some other keyword like that.

It is pretty easy to add "-character" keywords to as many of these titles as you want, rather than changing the existing keywords (or changing future "refererence-to-" keywords) to "-character" keywords. As I have said many times, the latter approach (apparently still your favored approach) is a lot more time-consuming to do, and also ends up eliminating relevant keywords and frustrating keyword searches. For titles with Christopher Columbus characters, the much better ultimate solution is to use both keywords "reference-to-christopher-columbus" and "christopher-columbus-character" side by side, rather than just one of these keywords. So if you are going to do this labor-intensive auditing, I would hope you would pursue the less time-consuming and least destructive approach. But it's your decision how to spend your time, and it's a free Internet.....

Having both the *-character" and *reference-to-" keywords is just unnecessary duplication and shoddy library/archive policy.  The "-character" keywords supercede the "reference-to-" keywords, making this duplication unnecessary and injurious to accurate research.

You seem to almost always want to take a supposedly expedient, "scorched earth" way that interfers with accurate auditing (even if it is time-consuming) and therein destroys or misrepresents much factual information

@bradley_kent​

In my opinion, you are the one who has adopted a scorched earth policy -- at least in this case -- by mass deleting relevant keywords that allow for keyword combination searches, rather than adding relevant keywords. 

In this case, to use your example, the "reference-to-christopher-columbus" keyword can and should serve as a master keyword for all titles having to do with Christopher Columbus, since only one keyword can serve that function (that's how keyword searches work). Here that master keyword can't be "christopher-columbus," which would violate the guidelines. 

You apparently prefer a scorched earth approach of destroying these keywords by deleting them from titles where they are relevant. You are scorching the earth one title at a time, in a slow burn sort of way. 

@bradley_kent​ 

I don't agree with you on this in some very specific cases. There are some movies/tv episode where a character is constantly referred to and then in a twist in the ending that character actually appears. In these cases, both reference-to and -character are appropriate.

@keyword_expert​ 

No. This is just silly to have "reference-to" for all titles having a character. Are you going to go in to all Batman episodes, movies and serials and add "reference-to-batman"? How in the world does that make any sense at all?

@adrian​ 

You are changing the subject. This thread is about real life people. Fictional characters that are part of a franchise are somewhat of a different story, since those types of keywords are allowed as standalone keywords. In that sense, "batman" may be allowed to serve as a master keyword, similar to how "star-wars" is allowed as a keyword, and "peanuts" is allowed as a keyword. 

For real life people, "reference-to-" keywords serve the same master-keyword function as "batman" and "peanuts." 

It's fine to also add "-character" keywords, but to proactively delete "reference-to-" keywords from these titles, which @bradley_kent is admitting he does as a routine course of action, is purging and destroying relevant data from the site.

I still adamantly maintain that, If a character APPEARS in a title, the "reference-to-" keyword is unnecessary and the "-character" keyword is needed.  Otherwise, you are 'opening up the flood gates' for "reference-to-" keywords EVERY time a character APPEARS, let alone just talked about or just has a picture/drawing/portrait in the title.

There must be some sort of distinction.  Otherwise, the keywords are just repetitive and redundant.

Jesus is "talked about" a lot in Ben-Hur with much anticipation, but, once he appears (and is portrayed by an actor), he is a "-character" keyword.  Happy. birthday, Jesus!

@bradley_kent​ 

And I still adamantly maintain that, If a character APPEARS in a title, both the "reference-to-" and "-character" keywords are valid.

You often say that you consider yourself a librarian. If that's the case, you should consider how keywords are actually used. 

In my opinion the single most important thing about keywords on IMDb is being able to quickly find patterns among titles. Here, for referring to "real life" people in keywords, the "reference-to-" keywords promote that use of keywords by serving as master keywords for titles that involve those people. 

With other types of keywords (e.g., "europe," "scientist," "explorer,") those keywords themselves serve as master keywords. But with real-life people, IMDb expressly forbids keywords limited to the names of people (e.g., "christopher-columbus"), so instead the only way to have a master keyword for that person is a keyword like "reference-to-christopher-columbus."

In this sense, the "reference-to-" prefixes are designed as IMDb's substitutes for linking to a person, place, or thing that otherwise appears in the database (other movies, people's names, character names, etc.).

And there are multiple ways where a person can be "referenced" in a movie -- there can be an indirect allusion to that person without ever saying their name or showing them on screen, they can be referred to in dialogue, they can appear on screen as a character or in archived footage, etc. Again, this includes titles where they are a character.

The bottom line is that the "reference-to-" keywords are the best way to track all titles that "reference" such people, and to do multi-keyword searches.

Using your Jesus example, I should be allowed to search for the following keywords in combination with each other, and to include all titles that include Jesus as a character as part of my search: "reference-to-jesus-christ", "religion", "miracle".

Sort by Popularity - Most Popular Movies and TV Shows tagged with keywords "reference-to-jesus-christ", "religion", "miracle"

You apparently believe that the "reference-to-jesus-christ" keyword should be eliminated (deleted) from titles where Jesus is a character, because you regard that keyword as extraneous. But every single time you delete a keyword in a situation like that, you are deleting relevant data and making it impossible to capture that title in a keyword combination search. 

Your approach interferes with keyword combination searches, since there is no way to do either/or keyword searches on IMDb (in other words, a search that would include either a "reference-to-" keyword or a "-character" keyword (or both keywords) in the same search). 

IMDb could solve this problem by allowing either/or keyword searches. But unless and until IMDb does that, your approach is eliminating relevant data and disrupting keyword combination searches.

It also doesn't work to do two separate searches: one with a "reference-to-" keyword and one with a "-character" keyword. That is not how keyword combination searches work. You literally will not be able to pick up all titles that use one or the other (or both) of these keywords, even if you do multiple searches. Again, that is because IMDb does not allow either/or keyword searches.

I  have "gotten used to" and become "somewhat immune" to your contrary rants to almost every position that I take regarding keywords.  While I agree with about 95% of your suggested mergers, I sure wish that you could be more objective about the positions of others.

It seems that you now want to "add" a "reference-to-" keyword to EVERY title that has a "-character" keyword.  This still seems totally unnecessary to me.  What IF the "-character" keyword represents the main character (protagonist)?  That is certainly more than a "reference-to-" situation.  Are "King of Kings" and "The Greatest Story Ever Told" and "Day of Triumph" and "The Passion of Chrst" to have a "reference-to-jesus-christ" keyword as well as the appropriate and more accurate "jesus-christ-character" keyword?

Oh, by the way, Merry Christ-mass! Or should that also be "reference-to-merry-christ-mass"?

@bradley_kent​ @adrian 

I started a new "idea" thread, linked below, to suggest that IMDb upgrade their keyword search options to fix the problems discussed in this thread. Unless and until these problems are fixed, "reference-to-" keywords will continue to be the best possible master keywords for titles involving real people. 

Advanced Keyword Searches: Allow multiple keywords to be selected from the Advanced Title Search screen, and allow "or" keyword searches