'Sensitive' property

[toby3d]

In the process of importing data from my Twitter and considering plans to import data from Mastodon to my website, I found out that there is no parameter related to "sensitivity" of the content in the attributes. The content published by the author on his site may not always be intended for everyone, for a variety of reasons, ranging from obvious "strong language" with 🍓 to chronicles from the scenes of controversial events. While for humans such publications can still be restricted by some kind of dialogue or other "yes, I really want to see this" mechanism, for machines and their owners it is now absolutely not obvious.

I propose to discuss several options for the name of the parameter and its values (binary and detailed) to decide and choose the appropriate.

p-sensitive / p-explicit / p-nsfw / etc.

Works similarly to p-rsvp, using one of the string case insensitive enum as value in data attribute:

• yes: The content potentially contains something unsafe to work with and is intended for a mature audience.
• no: The content is safe to work with and probably appropriate for any audience.

If the parameter is not specified at all, the default value is no. If specified without or with a different value, the value is assumed to be yes.

p-rating / other?

Not to be confused with the h-review evaluation. Works similarly to p-rsvp, using one of the string case insensitive enum as value in data attribute:

• explicit: The content clearly contains content that is not safe for work and is intended only for a mature audience.
• safe: The content clearly contains content that is completely safe to work with and intended for any audience.
• questionable: The rating has not yet been determined in favor of other values. Or the content contains content not intended for younger audiences, but acceptable for teens and older audiences.

If the parameter is not specified at all, specified without or with a different value, the value is assumed to be questionable.

Key with a non-normative value

As alternative parameters: specify a minimum age or age range, or specify a specific group. But in my opinion, this will quickly cause problems in their understanding from site to site. 🤔

---

I expect this property to apply not only to h-entry, but to any other unit of content like comments, reposts, likes and etc. As I mentioned earlier, the Twitter status API has an option possibly_sensitive: true/false, Mastodon has sensitive: true/false. In principle, a lot of public sites with relatively lenient usage rules expect that sooner or later they might publish unsafe content and strongly recommend that users mark suitable content themselves. This makes it possible to create "safe search" and other tools to filter content by this property.

[jgarber623]

This proposed property may be related to or a duplicate of #19.

[jalcine]

I'll leave a note to avoid the use of nsfw (Not Safe for Work) to align more to nsfk (Not Safe for Kids). It's more explicit and it implicitly reinforces that sex work is work versus the explicit banning of sex work as a concept.

[fluffy-critter]

"Safe for work" and "safe for kids" are both extremely subjective based on both personal and cultural standards. Having a more generic "content warning" is, IMO, a better umbrella concept.

[jalcine]

@fluffy-critter's completely right here.