Hacker News Reader

Comment by bosdev | Hacker News Reader

I dislike that the filters are strictly defined. What if I wanted to create a category for 'rape' or 'animal-cruelty'? I should be able to add those notes, and have them be silently ignored by unsupporting clients. Even better though, I don't see why the client wouldn't include those new categories in what can be filtered for this specific clip.

Edit: Also, it seems like the severity levels (low/medium/high) are arbitrary, meaning they will vary from film to film. It might make more sense to give categories sub-categories or tags. So for swear words, it would have the sub-categories of all the various words which might be used. For violence, it might have the tags of ('simulated', 'on screen', 'bloody'), etc. That way the viewer could decide what they want to see, and it would be somewhat consistent.

View on HN

Regarding the category-specific severity levels (instead of "low", "medium" and "high):

The three levels currently used are definitely not completely arbitrary. I'm sure we all can tell mild violence and really severe levels of violence apart. But I definitely see your point, nonetheless.

For "nudity", the different levels of "bare buttocks" versus "full (frontal) nudity" come to mind, for example. This caught be implemented in custom severity levels, of course.

But the major advantage of the three general levels is that the higher ones include the lower ones. That makes things as easy as choosing one severity level for each category.

If, for "nudity", you have the two levels suggested above (and potentially dozens of other levels), you have to provide a check box for every single level or "sub-category". That's far more complex, both from a consumer UX perspective and from the tagger experience.

Thanks for your feedback!

At present, the list of available categories and filters has been kept short to ensure simplicity.

We might think about allowing any custom categories, as you suggested. So we should definitely keep that idea in mind.

But let's assume you added "rape" as a category (which we thougth about as well): Then when do you add "rape" versus the existing "violence", "sex" or "nudity"? Do you have to add them all for most rape scenes? On the other hand, what do the consumers do? Do they have to tick dozens of boxes to ensure that no unwanted scenes are shown?

The more categories you add, the more blurred do the boundaries become.

I expect that enabling custom categories could lead to this being useful in ways that you'd never expect. For example, people with misophonia HATE the sound of people chewing loudly. It's not uncommon (the misophonia subreddit has 8,000 subscribers). I'd love to use your filter to mute the audio (and simultaneously turn on the subtitles, if possible) when there are misophonia triggers in movies. People with misophonia are already interested in movie warnings for this (e.g. the subreddit had several high profile warnings about the triggers in The Martian).

If misophonia is a possible useful application of custom categories, there might be many others.

Very good example, thank you!

So these small communities could really profit from such an open format, in case custom categories were added.

They will certainly not get this with the commercial providers of movie filtering.

> The more categories you add, the more blurred do the boundaries become.

I think that's kind of the point. There isn't a single coherent dependency for all cultures for all content.

To have a simplified single set is to place one (semi arbitrary) value system over another.

You could study the correlations and common concurrent uses, and maybe provide common suggestions, like Amazon does. But again, boundaries and weirdness.

But this is software that is intended to actually be used by humans. It would be a far worse user experience to prohibit the "sex" category but then be shown a rape scene because someone added "rape" as a custom category.

This concept really doesn't work without standardized categories. Of course, you could argue that more granularity should be added to the categories, and I don't necessarily disagree. You could even have a directed graph of categories, so that (for instance) "rape" could be a descendant of both "violence" and "sex" and thus rape scenes would be banned if the user banned either violence or sex.

Thanks for these interesting suggestions for implementation!

The only downside I could think of is that things could get a bit untransparent with such complex graphs. People could think: "Why did I miss this scene? I haven't even enabled that particular filter."

> There isn't a single coherent dependency for all cultures for all content.

Really an important point. Thanks!

Having this sort of control seems to be the only way for something like this to be useful across a wide range of audiences. My wife will watch people being murdered and maimed in the most horrifying of manners, but the second an animal comes to harm she'll get up and walk out of the room (and then blame me for making her watch it).

"Violence" doesn't cover her specific use case, so without the ability to narrow that to "animal cruelty", it wouldn't address our own requirements.

Thanks, that's definitely a valid concern! You're both right in that we should perhaps change the severity schema to category-specific keywords instead of general levels.

Tags work best! Like on SO

Why not multiple tags then? Rape is a specific instance of both sex and violence (and potentially includes nudity). If someone wanted all violence and/or sex filtered out then rape would be as well. But if they had no issue with violence in general, or sex in general, but specifically an issue with rape, then they could go to that level of specificity.

Two feature requests:

1) Add a "product placement" category.

2) Add support for a custom mask instead of just blanking out video/audio. This way, I can see that the character is drinking a delicious soda, but I don't need to know the exact kind.

Thank you! Regarding (1), we should definitely add this. Regarding (2), there are two reasons why this might not be a good idea: First, it would be hard to tag. Second, it's not supported by the three formats that the project currently "transpiles" to.

> Then when do you add "rape" versus the existing "violence", "sex" or "nudity"? Do you have to add them all for most rape scenes?

Rape is sex without consent, big difference between that and violent (consensual) sex.

"rape" might also include some frat boys joking about rape, which might really upset a rape victim watching it.

> Rape is sex without consent, big difference between that and violent (consensual) sex.

We don't have to discuss this. Everybody should know this and agree.

But the point was that rape scenes often include sex or nudity. So if you tagged a scene 'rape', you would often have to add the other related tags as well. This is because somebody might have banned 'nudity' only.

> We don't have to discuss this. Everybody should know this and agree.

The comment I was replying to, implied that you could just have violence+sex+nudity categories instead:

when do you add "rape" versus the existing "violence", "sex" or "nudity"?

Isn't the latter more metarape than rape?

A "rape" category doesn't have to include scenes of the act. Someone who doesn't want to see anything to do with rape, would not want to hear rape jokes.

It might make sense to add a (textual) description of the severity levels to the format. Then, the viewer can check that description to decide what level they are comfortable with.

The obvious drawback of that approach (shared with tags) is that the description needs to be in some language. This clearly hurts internationalization efforts.

Thanks! Thought about that, too. Some example:

nudity/low: bare buttocks

nudity/high: full (frontal) nudity

But this is harder (maybe even impossible) in other categories. What if somebody sees "throat cut" as a "high" level of violence and a killed animal as "low" level, while someone else sees it the other way round?

Reply on news.ycombinator.com