Keywords now displayed on the IMDb iOS and Android apps

@Peter_pbn​ Great suggestion, thank you. We will pass this on for consideration.

Here's the GWTW contribution, in which, interesting, three of the Plot Timeframe category submissions were not accepted.

Now, EVERY SINGLE KEYWORD will need to be audited and corrected to place EVERY SINGLE KEYWORD into a category!

I don't see anyone asking you to add categories to all existing keywords.

Keywords without a category are displayed alongside those marked Plot Detail, so I wouldn't spend time adding that category.

One must change to "Correct" on EVERY SINGLE KEYWORD before one can even begin correcting them.  (There should be some way to automatically do this rather than the time-consuming manual, keyword-by-keywird way!)

There is.

I had to enter the "Other" option for the following subgenres

By using "Other" you will keep the keywords away from the main title page, so I would suggest leaving them alone.

Thanks for your feedback

You are right that it will take some time to backfill keywords into the correct category. For now we have defaulted all keywords to "plot-detail" - the category that most keywords apply to. We then migrated all valid subgenre and plot timeframe values to their respective categories, and also backfilled the 1,000 most common keywords in the catalog to their appropriate categories. We definitely appreciate that it may seem daunting to migrate the long tail of remaining keywords. Please do share any keywords that need migrating to new categories on this thread and we will look into moving them. Please bear with us - you are right that it will take time.

To your point about plot timeframe keywords - we are in the process of removing this prefix from all items in the catalog and relying on the category field instead. This is in part a response to the feedback we received about the prefix from this community. Please bear with us while we process through these. In the meantime, we have made updates to the Help guide and contribution interface to reflect this new policy.

Thank you for pointing out those candidate subgenre options...

Regarding the following suggestions, we are still opting to tie subgenres to genres. "Epic" is one of the rarer and more extreme examples, where we have created it as an option for most genres, e.g. epic-action, epic-adventure, epic-fantasy. We realize there are drawbacks to this approach, but have also found this structure to be useful:

• epic: Epic is currently an option for most genres, e.g. epic-action, epic-adventure, epic-fantasy.
• melodrama: This is a potential subgenre option for drama. 
• farce: We opted to capture farce as "slapstick" for now. We appreciate this isn't always entirely accurate, but feel that slapstick would be more familiar to the broader entertainment audience.
• opera: We have the option "operetta-musical". We could also include opera under the music genre if that sounds right to you?
• neo-noir: This is a tricky one that we discussed a bit and don't yet have a solution for. The Film Noir genre guideline states that the neo-noir keyword should only be submitted as a keyword for titles that do not fit all criteria of the Film Noir genre. We are curious to explore solutions.

For the other candidates:

• political-drama: This is already included in the subgenre options.
• romantic-drama: In response to previous feedback from this community, we have opted to capture these scenarios with combinations of genres (Romance+Drama genres instead of romantic-drama subgenre) going forward, instead of duplicating the same data under subgenre keywords.
• tearjerker-womance: We have this option as "tearjerker-romance".
• tragic-romance: This is already included in the subgenre options.
• wartime-romance: We would capture this with the genre combination of War+Romance.

Thank you again for your feedback and for bearing with us! We look forward to discussing further.

@bradley_kent​ Those 3 plot-timeframe items are now approved - thank you for flagging those.

Just to say that biographical-documentary and history-documentary will also be removed from the subgenre options in the Help guidelines. The update just hasn't reached the Help site yet. We will instead rely on combinations of those genres, per previous feedback from this community.

Thanks

Farce and slapstick are two different kinds of “low comedy.”  A farce MAY include slapstick comedy, but not necessarily so.

Oxford Dictionary definition of farce:

a comic dramatic work using buffoonery and horseplay and typically including crude characterization and ludicrously improbable situations.

Definition of slapstick:

the type of humour that is based on simple actions, for example people hitting each other, falling down, etc.

“slapstick-comedy” is a better keyword than “slapstick” since “slapstick” also refers to the literal object (a slap board) used to produce a slapping sound, most commonly used in Punch and Judy puppet shows, but also in other “low” forms on comedies, as in Scaramouche (1952).

@mike​ 

Looks like some good improvements have been made.

I do question the idea that each subgenre keyword should have to fit underneath a specific genre. I understand why this might be your goal, but in reality a lot of subgenre keywords don't work that way.

Here are just a few random "subgenre" keywords (for lack of a better term) that don't necessarily fit with one specific genre:

giallo (1913 titles)  

exploitation (2329 titles)

exploitation-film (995 titles)

sexploitation (1483 titles)

sexploitation-film (922 titles)

italian-exploitation (317 titles)

psychotronic (24 titles) (whoah! what happened to this one? in my opinion all of the various "psychotronic-" keywords should be merged into "psychotronic," rather than the other way around. but that thinking may outtdated given the new changes to the site)

psychotronic-film (8894 titles)

psychotronic-series (1251 titles)

social-guidance (215 titles)

nunsploitation (138 titles)

bikersploitation (45 titles)

teensploitation (267 titles)

drugsploitation (47 titles)

creature-feature (1247 titles)

rape-and-revenge (1369 titles)

women-in-prison (289 titles)

grindhouse-film (2416 titles)

slimehouse (1130 titles)

shrekcore (144 titles)

b-movie (4279 titles)

z-movie (856 titles)

trash-movie (864 titles)  (this one should really be merged into "z-movie")

euro-trash (852 titles)

sleazy-giallo (251 titles)

candy-cinema (252 titles)

lgbt-interest (265 titles) (some of these lgbt keywords could potentially be merged)

lgbt-cinema (389 titles)

queer-cinema (502 titles)

lgbt-film (51 titles)

lgbtq-film (29 titles)

surrealism (10982 titles)

avant-garde (2848 titles)

experimental-film (4740 titles)

experimental (3846 titles)

Most of those are probably not even properly classified as subgenres in the first place. That explains why it's hard to pigeonhole specific keywords from this list underneath a single genre.

I don't regard wartime-romance (106 titles) as a subgenre keyword either -- or at least not on some titles where it has been used. In fact, I find "wartime-romance" to be a good plot keyword for titles that do not qualify for both the War and Romance genres. 

@mike​ 

• neo-noir: This is a tricky one that we discussed a bit and don't yet have a solution for. The  Film Noir genre guideline states that the neo-noir keyword should only be submitted as a keyword for titles that do not fit all criteria of the Film Noir genre. We are curious to explore solutions.

This is only a problem if IMDb staff continue to insist that all subgenre keywords must contain a main-genre suffix and must be grouped under a single main genre.

It has never made much sense to handle subgenre keywords that way, and in fact it is borderline dysfunctional because it excludes dozens of established subgenres and quasi-subgenres.

I see absolutely no problem with treating "neo-noir" as a Subgenre keyword. Furthermore, this keyword doesn't need to be grouped under the Film Noir genre just because they both contain the word "noir." It's not like there is a genre hierarchy/mapping on the website that forces each subgenre to be grouped under only one genre. Yet staff are treating subgenres this way. I don't even recall ever seeing a justification or explanation of why this approach has been adopted or deemed necessary.

Finally, I see that IMDb staff are categorizing the keyword "alternate-history" as a Subgenre keyword, even though this keyword by definition does not fit underneath the History genre. I am completely fine with this; "alternate-history" should be a subgenre, even though all titles that have the "alternate-history" keyword will not qualify for the History genre.

The staff should treat "neo-noir" the same way it treats "alternate-history," by allowing both keywords to be designated as subgenres. 

@keyword_expert​ 

I don't even recall ever seeing a justification or explanation of why this approach has been adopted or deemed necessary.

The usual definition of subgenre is a subdivision of a genre. Thus I don't think it is hard to see where the idea of attaching subgenres to specific genres comes from.

@Peter_pbn

I wasn't asking where the idea came from, but rather why this approach has been deemed necessary. ​

Is there more to it than just a broad conceptual understanding of what a subgenre is? That is more of a philosophical/etymological question than a programming/data/website question.

We need to hear an explanation from staff. 

@bradley_kent​ 

mike is not suggesting that opera and operetta are the same thing, but that opera could be listed as opera-music like operetta is listed as operetta-musical.

opera-music is slightly problematic because opera is a performance form as much as a style of music. The list of subgenres already has classical-music, which would include opera music.

I don't agree that opera is a subgenre of musical.

Opera could also be added as a genre.

Such music OUTSIDE performance is usually called "classical-music" or "operatic-music."

In a cursory way, I keep coming across subgenres that now can only be miscategorized as "Other." Isn't "anime" a subgenre of Animation?  Isn't "cinema-verte" a subgenre?  "mockumentary?" "neo-noir," a subgenre of Film Noir?  (It's now only an "Other.") "neo-screwball-comedy," a subgenre of Screwball Comedy? (I did a gargantuan separation of such titles a few years ago that took a lot of time and conscientious persistence,)

And, then there are many supposed subgenres that probably have ONLY a few titles, like "lo-fi-sci-fi," "fake-documentary" (to be merged with "mockumentary," or are they two separate subgenres?). Etc. Etc. Etc.

Before this program was instituted, one of the first steps that should have been undertaken was to identify all valid subgenres that already exist in the database, and not just to rely on the new, and occasionally dubious subgenres.  They are only a small part of valid subgenres that already exist.  (Isn't "tragedy" a subgenre of Drama?)

As to OTHER choices in the Category drop-down menu, there needs to be something like "production-techniques" to cover things like "time-lapse-photography," "slow-motion-scene," "cgi," ""montage," "long-take," etc., and perhaps something like "literary sources" to include keywords like "based-on-novel," "based-on-comic-book," "french-literature-on-screen," etc.  These may not be the best names for a category.  I'm sure that someone else can come up with better category names to cover these "non-plot" keywords..

Anime was relisted as anime-animation, mockumentary as mockumentary-comedy, tragedy as tragedy-drama. The old keywords remain, but I doubt we have arrived at the final result just yet.

"-animation," "-comedy" and "-drama" are all redundant.  Again, just one of the problems with the "new" subgenres.

Thanks all for the great suggestions!

Regarding some of the subgenre suggestions (we will follow up on others):

• We have now enabled farce-comedy as a subgenre in the contribution interface. It will also be added to the Help article.
• Giallo is currently categorized as a subgenre of thriller.
• We haven’t made any changes to psychotronic keywords as part of this launch (they are not currently classified under the subgenre category), but please let us know if there are specific keywords that should be merged here.
• You are right that the genre suffix for subgenre keywords is awkward. You are also right that some subgenres don’t have a strong correlation to specific genres (e.g. “epic-”). We still see this as a worthwhile trade-off given the consistency and structure it provides. However, this launch of a contributable category field could provide us more freedom to evolve the format in future. We will continue to evaluate this.

Regarding the category suggestions:

• We had been thinking about exploitation, avant-garde etc. One option is to categorize them under something like "Niche". We are curious to hear opinions on this.
• Production techniques, literary sources and geographical settings are great ideas. We had also been thinking about some of these.
• Other potential categories include Mood (e.g. dark/uplifting), Brand/Franchise (e.g. Marvel Cinematic Universe), Plot Environment (e.g. desert/space), Occasion (e.g. Thanksgiving/Halloween).

There are lots of opportunities to make keywords more digestible with further categorization. We also recognize that there are trade-offs as we add more categories - not least the work to categorize them. We plan to learn a bit more from the limited set of categories that we have launched before exploring these new ones.

In the meantime, it is fine for such keywords to live under the categories of Plot Details or Other. In general, if it describes a detail about the plot, then it should be under "Plot Detail". If it doesn't fit into any of the supported categories, then it should go under "Other". As mentioned, "Other" keywords are only shown on the subpage in the apps, although as before they can appear on the main or subpage on web (which doesn't have any categorization yet).

Again, the use of a genre as a suffix is , yes, awkward, but, more importantly, repetitive, redundant,  and absolutely unnecessary.

@mike​

Regarding the "psychotronic" keyword, I was not intending to imply that staff made any changes. It looks to me like over the past week, a contributor has either mass-manual-deleted that keyword or converted it to "psychotronic-film," "psychotronic-series," etc.

I dislike these keywords anyway, but I had actually been contemplating proposing the exact opposite idea: merging all "psychotronic-" keywords into the "psychotronic" keyword.

For those who use this keyword in keyword-combination searches (and I know for a fact people do exactly that, since I have seen people post those searches online), it would be better to simply use the keyword "psychotronic" rather than "psychotronic-series," "psychotronic-film," etc., so that more keywords can be picked up in combination with each other in a single search. 

Besides, the "-series," "-film," etc. suffixes of these keywords replicates other information in IMDb that can be filtered via searches. In other words, if all these keywords were merged into "psychotronic," then users could filter the results by feature films, TV series, etc. -- that part does not belong in the keyword itself.

In conclusion, I would still recommend mass merging all "psychotronic-" keywords into the single keyword "psychotronic." 

There is no need for staff to take that action now, since I can include it in a future public list. 

In the meantime, if anyone opposes my idea, it would be good to know the reasons why. 

The "-(feature)film," "-short," "-series," "-episode" suffixes are valuable.  I have searched them when programing films for a senior center.  Years ago, when teaching film, I would have appreciated them since time constraints always affected the films that I was selecting.

Also, they are just valuable for recording film history.  Perhaps "-web-series" and "-podcast" need to be treated in the same way.

Such keywords are a much easier search tool that what has been suggested.

@bradley_kent​ 

When you search for "psychotronic-film," "psychotronic-short," etc., why can't you just search for "psychotronic" and then if want to filter by Short or feature film or whatever, just filter your search results in that way (using the data already available on the IMDb site)?

The filtering should not be done within the keywords themselves. Doing it that way bifurcates the keywords and short-circuits keyword combination searches, unless the filtered keywords and general keywords (e.g., "psychotronic" and "psychotronic-series" are retained together, side by side on the same title).

The overall best solution is to just use the single keyword "psychotronic."

Imagine if every single keyword involving subgenres and styles were compartmentalized like you have done with "psychotronic." We would end up with "bikersploitation-series," "bikersploitation-short," "bikersploitation-film," etc. keywords, and then repeat all those combinations for every single subgenre/style keyword.  All of that is completely unnecessary.

But more importantly, it interferes with keyword combination searches. This is the main problem here.

For example, if you want to search for the keywords "psychotronic," "ghost," and "woods" together, you should be able to do that search to get all the applicable titles, and then if you want to filter it out by TV series, film, short, or whatever, you can add that to the search. But if you don't want to do that filtering, you should be able to search for these keywords together in a single search.

With your approach, you would have to do multiple searches, and there is no single way to combine all of those multiple searches into a single search (or a link that can be shared with others).

In other matters (for example, genres), you have been a big proponent of not using keywords to duplicate data that exists elsewhere. Why are you advocating for the duplication of data within keywords here?

As a recent example, IMDb staff recently concluded that keywords shouldn't contain "tv-movie" within the keywords.

If "tv-movie" within a keyword is not allowed, why should "-series" be allowed within a keyword?

I really hope you don't start manually editing all the other subgenre/style keywords to break them down by series, film, short, episode, etc. That would be a lot of wasted time that would, in the end, only be counterproductive.

It all seems to depend on HOW you want to do your search, whether exclusively using keywords or combining keywords with genres.  Also, some "short" films have no running time, no "short" designation in the keywords, and rely only on the genre designation.

Also, "psychotronic" is a standalone adjective.  Standalone adjectives should generally be avoided In keywords since one cannot then know what that adjective is specifying.

@bradley_kent​ 

I'm mainly talking about multi-keyword-combination searches. That's the main problem created with your approach of bifurcating these keywords.

To solve this problem, there are two potential solutions. First, all titles that have keywords like "psychotronic-series" and the like could also have the keyword "psychotronic."

The second, preferred option is that there should not be any keywords like "psychotronic-series" or "psychotronic-film" at all, and instead only one keyword ("psychotronic") is needed.

Sure, "psychotronic" is an adjective. But it's also an understood genre/style that will be recognized by people who are familiar with it -- and especially if it is designated as a subgenre keyword or whatever IMDb's new system is for those types of keywords. It's the same with keywords like "avant-garde," "experimental," etc. None of those keywords should be banned just because they are adjectives.

@bradley_kent​

BTW, "giallo" is also a standalone adjective, and it's not even in English, so it has two strikes against it as a keyword. Would that mean you are also going to slice and dice the keyword "giallo" up into "giallo-film," "giallo-short," etc.?

Your approach in favor of keywords like "psychotronic-short" seems very much at odds with your approach against genre-infused keywords.

There is no keyword "short." Nor should there be any keyword like "psychotronic-short," "independent-short," "experimental-short," etc. All of those keywords should be merged into their counterparts "psychotronic," "independent-film," "experimental," etc.  One day I will likely propose a set of mergers like this in a future list.

@mike​ 

• You are right that the genre suffix for subgenre keywords is awkward. You are also right that some subgenres don’t have a strong correlation to specific genres (e.g. “epic-”). We still see this as a worthwhile trade-off given the consistency and structure it provides. However, this launch of a contributable category field could provide us more freedom to evolve the format in future. We will continue to evaluate this.

As I understand it, IMDb is currently only accepting keywords as Subgenre keywords if (1) the keyword includes a genre as a suffix and (2) the keyword has been added to a pre-approved list by IMDb staff.

I would encourage the staff to expand the list of pre-approved Subgenre keywords. As you acknowledge in the sentence I have bolded above, now that we have the ability to submit a keyword as a Subgenre keyword, that should obviate the need for including a genre suffix in all subgenre keywords.

The following keywords should be added to the list of pre-approved Subgenre keywords (and should be retroactively designated as Subgenre keywords where they have already been applied):

giallo (1913 titles)  

exploitation (2329 titles)

exploitation-film (995 titles)

sexploitation (1483 titles)

sexploitation-film (922 titles)

italian-exploitation (317 titles)

psychotronic (24 titles) 

social-guidance (215 titles)

nunsploitation (138 titles)

bikersploitation (45 titles)

teensploitation (267 titles)

drugsploitation (47 titles)

creature-feature (1247 titles)

rape-and-revenge (1369 titles)

women-in-prison (289 titles)

grindhouse-film (2416 titles)

slimehouse (1130 titles)

shrekcore (144 titles)

b-movie (4279 titles)

z-movie (856 titles)

euro-trash (852 titles)

sleazy-giallo (251 titles)

candy-cinema (252 titles)

gay-cinema (414 titles)

lgbt-interest (265 titles) 

lgbt-cinema (389 titles)

queer-cinema (502 titles)

lgbt-film (51 titles)

lgbtq-film (29 titles)

surrealism (10982 titles)

avant-garde (2848 titles)

experimental-film (4740 titles)

experimental (3846 titles)

chop-socky (517 titles)

kung-fu-classic (309 titles)

kung-fu-comedy (22 titles)

mockumentary (2433 titles)

hillbilly-horror (165 titles)

police-drama (17 titles)

period-piece (662 titles)

alternate-history (442 titles)

postmodern (330 titles)

docufiction (38 titles)

docusoap (128 titles)

cult-film (5073 titles)

cult-classic (358 titles)

mature-animation (1142 titles)

anime (6064 titles)

buddy-movie (242 titles)

satire (6400 titles)

shark-feature (129 titles)

sharksploitation (117 titles)

sex-horror (221 titles)

erotic-horror (78 titles)

urban-horror (71 titles)

lgbt-horror (57 titles)

retro-horror (106 titles)

gothic-horror (297 titles)

horror-for-children (205 titles)

australian-horror (315 titles)

japanese-horror (247 titles)

american-horror (210 titles)

christmas-horror (164 titles)

classic-horror (147 titles)

british-horror (107 titles)

arthouse-horror (57 titles)

neo-noir (4033 titles)

softcore (14169 titles)

sitcom (2385 titles)

eurospy (225 titles)

@keyword_expert​ 

forest-adventure was removed from the subgenre list in July, as I documented.

@Peter_pbn​ Seems like it should go back on the list. It definitely shouldn't be categorized as a Plot Detail.