I have been thinking about this sort of thing too. I really like the idea of a social-network type thing with an emphasis on transparency and personal control rather than black-box feed curation.
I want to re-emphasize what I see as your main two points:
A) People are the center of their own webs of trust, meaning that they have complete control over the initial seeds of trust which are used to calculate extended-trust-families for them.
B) People are not forced to be uni-faceted, meaning (B1) there is not just a single dimension of trust, but rather many kinds of trust (EG represented by a tagging system); (B2) people’s individual content is not automatically assumed to fall into the same type all the time—for example, someone who is trusted by a large number of people for good musical analysis should be able to start writing about mushrooms and herbs without souring their existing trust.
I strongly expect attempted implementations of this sort of thing to drop one of these points, simply because it’s a lot of complexity to manage (both on the back end and as a matter of UI).
I want to slightly re-work your “concept” sketch:
Presence/Article: As you have said.
Tag: Basically as you’ve said, but we can also consider quantitative judgements (like tag votes, or star ratings). This also allows for explicit downvoting.
Tag Trust: Just like a tag, but applied to presences instead of articles. The meaning is definitely not “this person produces this type of content” (see point B2); rather, it means I trust this person to tag this type of content (and I trust them to tag other people who tag this type of content, and so on recursively).
Important things about this:
Tag trust is about trusting a person as a curator, not inherently about trusting them as a producer of that type of content at all.
Tagging is a lot like “sharing” on facebook/twitter/etc.
When people produce content, they put some initial tags on it. They’re self-curating (or self-sharing, if you will). This is what lets the musical analyst also talk about mushrooms and herbs, without creating an alt account. (I have nothing against creating alt accounts to do this, too, but tags seem more flexible, since you can tag content with multiple tags.)
There might also be a concept of following a presence. For example, things could be auto-tagged with usernames, so that you can follow a user’s tag to see all their stuff.
I asserted that tag trust means two things:
I trust this person to curate this type of content, IE, apply the tag well.
I trust this person to apply the corresponding tag trust well, too.
Realistically, there could be a need to split this into two separate concepts. For example, some people might play a strong “filtering” role—let’s say someone reads political theory very broadly, but tags it very selectively. This means I don’t trust their tag trust very much, because they apply it really broadly in order to get lots of reading material. However, I trust their actual tags on the subject, because they apply those very selectively.
Not sure how to handle that without getting real complicated, though.
I really appreciate having someone else who gets it summarize the concept with their own emphases added. Extremely helpful for figuring out how to communicate the thing.
Mm, so the Tag Trust way of presenting things would boil down like… users would implicitly have a presence associated with each tag, that resides in a web implicitly associated with each tag. (This is what -taste webs were about, btw, but it hadn’t occurred to me til now that the existence of presences in those should always be implicit.)
Maybe users should be introduced to those sorts of single-tag webs before being told that using a web to curate more than one tag at the same time is possible, to prevent web overloading from breaking out sooner than it needs to. If web overloading only came from intermediate users that would be really good. Subcultures only, no platform superculture.
There might also be a concept of following a presence.
It might be beneficial to implement following as just endorsing them, in some way, from one of your presences, as endorsements contribute structure that other people can use (if they want to). Even if they’re not very wise endorsements, they might be useful for some things while the web is growing. It’s a way of getting people to start building on the structure without even needing to explain how it works.
Distinguishing the personal quality itself from the ability to recognize the quality in others
It seems like it would be good to have a type of trust relation that ignore the endorsed’s endorsements, stopping the query frontier from propagating beyond them. The person you describe is sort of using the platform incorrectly though, I’m not sure it would turn out to be necessary.
If they know that their outputs are higher quality than their inputs, they should separate them (we should make it easy for them to separate them). If they care enough to make sure they only tag the really good stuff, wouldn’t they then care enough to tag them with a curation presence (that endorses a different set of presences (or nobody) to their research presence) so that people can just subscribe to that instead.
Webs of similar topics should generally spring up along a continuum from decreasing breadth to increasing quality (for example, game releases are taken by enthusiastic consumers to the web game recommendations. Most people would only pay attention to game recommendations).
We could properly define “Curators” as users who have active presences in at least two separate webs, where the webs have the same overall topic, but one is more discerning.
Maybe that should be part of the UI. If two webs are considered to be on the same refinement continuum, it should be made visible, and it should be made easy for curators to tag articles along their different presences in the same continuum with a single action.
Whenever you’re browsing under a tag, posts would have a button that lets you agree with the tag, applying it using the presence you’re looking at it through. Maybe if you’re browsing through a curator input presence that button does an additional thing, tagging it not just with the presence you’re browsing with, but also with its corresponding curation presence in the next refinement level, so it would be more “promote” or “escalate”. (related: if you are one of the pesky users who insists on using the general taste web, that button would tag it on your general taste presence too)
I have been thinking about this sort of thing too. I really like the idea of a social-network type thing with an emphasis on transparency and personal control rather than black-box feed curation.
I want to re-emphasize what I see as your main two points:
A) People are the center of their own webs of trust, meaning that they have complete control over the initial seeds of trust which are used to calculate extended-trust-families for them.
B) People are not forced to be uni-faceted, meaning (B1) there is not just a single dimension of trust, but rather many kinds of trust (EG represented by a tagging system); (B2) people’s individual content is not automatically assumed to fall into the same type all the time—for example, someone who is trusted by a large number of people for good musical analysis should be able to start writing about mushrooms and herbs without souring their existing trust.
I strongly expect attempted implementations of this sort of thing to drop one of these points, simply because it’s a lot of complexity to manage (both on the back end and as a matter of UI).
I want to slightly re-work your “concept” sketch:
Presence/Article: As you have said.
Tag: Basically as you’ve said, but we can also consider quantitative judgements (like tag votes, or star ratings). This also allows for explicit downvoting.
Tag Trust: Just like a tag, but applied to presences instead of articles. The meaning is definitely not “this person produces this type of content” (see point B2); rather, it means I trust this person to tag this type of content (and I trust them to tag other people who tag this type of content, and so on recursively).
Important things about this:
Tag trust is about trusting a person as a curator, not inherently about trusting them as a producer of that type of content at all.
Tagging is a lot like “sharing” on facebook/twitter/etc.
When people produce content, they put some initial tags on it. They’re self-curating (or self-sharing, if you will). This is what lets the musical analyst also talk about mushrooms and herbs, without creating an alt account. (I have nothing against creating alt accounts to do this, too, but tags seem more flexible, since you can tag content with multiple tags.)
There might also be a concept of following a presence. For example, things could be auto-tagged with usernames, so that you can follow a user’s tag to see all their stuff.
I asserted that tag trust means two things:
I trust this person to curate this type of content, IE, apply the tag well.
I trust this person to apply the corresponding tag trust well, too.
Realistically, there could be a need to split this into two separate concepts. For example, some people might play a strong “filtering” role—let’s say someone reads political theory very broadly, but tags it very selectively. This means I don’t trust their tag trust very much, because they apply it really broadly in order to get lots of reading material. However, I trust their actual tags on the subject, because they apply those very selectively.
Not sure how to handle that without getting real complicated, though.
I really appreciate having someone else who gets it summarize the concept with their own emphases added. Extremely helpful for figuring out how to communicate the thing.
Mm, so the Tag Trust way of presenting things would boil down like… users would implicitly have a presence associated with each tag, that resides in a web implicitly associated with each tag. (This is what -taste webs were about, btw, but it hadn’t occurred to me til now that the existence of presences in those should always be implicit.)
Maybe users should be introduced to those sorts of single-tag webs before being told that using a web to curate more than one tag at the same time is possible, to prevent web overloading from breaking out sooner than it needs to. If web overloading only came from intermediate users that would be really good. Subcultures only, no platform superculture.
It might be beneficial to implement following as just endorsing them, in some way, from one of your presences, as endorsements contribute structure that other people can use (if they want to). Even if they’re not very wise endorsements, they might be useful for some things while the web is growing. It’s a way of getting people to start building on the structure without even needing to explain how it works.
Distinguishing the personal quality itself from the ability to recognize the quality in others
It seems like it would be good to have a type of trust relation that ignore the endorsed’s endorsements, stopping the query frontier from propagating beyond them. The person you describe is sort of using the platform incorrectly though, I’m not sure it would turn out to be necessary.
If they know that their outputs are higher quality than their inputs, they should separate them (we should make it easy for them to separate them). If they care enough to make sure they only tag the really good stuff, wouldn’t they then care enough to tag them with a curation presence (that endorses a different set of presences (or nobody) to their research presence) so that people can just subscribe to that instead.
Webs of similar topics should generally spring up along a continuum from decreasing breadth to increasing quality (for example, game releases are taken by enthusiastic consumers to the web game recommendations. Most people would only pay attention to game recommendations).
We could properly define “Curators” as users who have active presences in at least two separate webs, where the webs have the same overall topic, but one is more discerning.
Maybe that should be part of the UI. If two webs are considered to be on the same refinement continuum, it should be made visible, and it should be made easy for curators to tag articles along their different presences in the same continuum with a single action.
Whenever you’re browsing under a tag, posts would have a button that lets you agree with the tag, applying it using the presence you’re looking at it through. Maybe if you’re browsing through a curator input presence that button does an additional thing, tagging it not just with the presence you’re browsing with, but also with its corresponding curation presence in the next refinement level, so it would be more “promote” or “escalate”. (related: if you are one of the pesky users who insists on using the general taste web, that button would tag it on your general taste presence too)