4 ms·

Actually, what properties would such a place have? I keep thinking that a defining problem of the present day is the absence of such a platform. I'm even willi

by Folcon 8y ago

Actually, what properties would such a place have? I keep thinking that a defining problem of the present day is the absence of such a platform. I'm even willing to dig in and help put something together. I wonder if an ask HN around: - How people would like to interact with such a system. - What it would need to have/do to be useful. I mean is there even a good format for such a thing? Debates for example are explicitly two sided, but exploring a problem rarely is. You need to track what's been tried, why it might have failed, see if it's worth trying again, allow people to suggest amendments which might help. There's so much there.

TeMPOraL 8y ago
A bunch of ideas: 1) It would need to have good mechanisms for structuring knowledge and discovery (and deduplication) of previous points. Instead of time-based approach, where people read and respond to streams of comments, it should be easy to review past discussions and maybe add updates based on new information or situations that develop. Structuring and discovery is hard. I spent some time on Kialo[0], which is a "debate platform" that encourages tree-structured arguments. For any stated problem, people post arguments for and against, and each of those arguments can be recursively subdivided into for and against points, ad infinitum. It's a cute concept, but my impression was that the tree structure isn't good enough fit for the problem - frequently, you'd find the same points repeated at many different levels. A point "X" can be found simultaneously as a top-level "against" argument, and as a support argument three levels deep inside a top-level "for" argument, etc. The experience reinforced my feelings that problem spaces tend to form cyclic graphs. 2) It would need to have some mechanisms to encourage proper discourse standards. Moderation is one of them, but also some structural things more focused than just free-form comment textbox. 3) Ultimately, the goal of a platform should be to flesh out the problem space. The result should be an explorable artifact. This would help keep people on the same page, as well as get new participants up to speed (and inform non-participants who just want a honest overview of a problem they're interested in). I have no clue how such system should be built, but I'd love to see and play with any experiments in this space. -- [0] - https://www.kialo.com/tour/ https://www.kialo.com/tour/
- Folcon 8y ago
  Ok that's interesting, you've clearly done a lot in this space already, are you still working in this space and would you be up for a quick discussion about it? I was actually thinking about it from a different direction. Allow people to enter free text and then aggressively parse it out into a structure. Make the structure visible and then work with the structure. But allow the free text to stand as metadata. Allow the person writing to highlight/markup blocks in the main text that support their argument. The structure I was thinking of as a starting point was a list of clauses. You could then surface people stating a clause, saying that it's sometimes true or false, disagreeing with it, and detect it across the nodes in the graph making discovery a bit easier. Or even taking a clause and using it as a starting point for a new discussion. The other thing I think that's necessary is what kind of environment you foster. Something like changemyview[0][1] with the delta system [2] comes to mind. Simply speaking a person awards delta's when they've had their mind changed in any meaningful way by a person, and they're required to give a short explanation or statement about why they awarded it. -- - [0] https://changemyview.net/ https://changemyview.net/ - [1] https://www.reddit.com/r/changemyview/ https://www.reddit.com/r/changemyview/ - [2] https://www.reddit.com/r/changemyview/wiki/deltasystem https://www.reddit.com/r/changemyview/wiki/deltasystem
  - TeMPOraL 8y ago
    > Ok that's interesting, you've clearly done a lot in this space already, are you still working in this space and would you be up for a quick discussion about it? Sure! Feel free to mail me (address in the profile). I wouldn't say I have done a lot in this space - I am just interested in it, and try to pay attention to any new ways to structure discussions and thoughts that helps people reach agreement. -- Your concept is very interesting, and indeed a completely different direction than what I was thinking. You seem to want to start from making free-text comments more machine-understandable, so that they can be further analyzed and grouped with similar points from other people? This seems it could mesh with what I wrote in the previous comment - I want to get a graph of ideas, and you just suggested an approach to generating that graph.
    - 8y ago

by Folcon 8y ago

TeMPOraL 8y ago

A bunch of ideas: 1) It would need to have good mechanisms for structuring knowledge and discovery (and deduplication) of previous points. Instead of time-based approach, where people read and respond to streams of comments, it should be easy to review past discussions and maybe add updates based on new information or situations that develop. Structuring and discovery is hard. I spent some time on Kialo[0], which is a "debate platform" that encourages tree-structured arguments. For any stated problem, people post arguments for and against, and each of those arguments can be recursively subdivided into for and against points, ad infinitum. It's a cute concept, but my impression was that the tree structure isn't good enough fit for the problem - frequently, you'd find the same points repeated at many different levels. A point "X" can be found simultaneously as a top-level "against" argument, and as a support argument three levels deep inside a top-level "for" argument, etc. The experience reinforced my feelings that problem spaces tend to form cyclic graphs. 2) It would need to have some mechanisms to encourage proper discourse standards. Moderation is one of them, but also some structural things more focused than just free-form comment textbox. 3) Ultimately, the goal of a platform should be to flesh out the problem space. The result should be an explorable artifact. This would help keep people on the same page, as well as get new participants up to speed (and inform non-participants who just want a honest overview of a problem they're interested in). I have no clue how such system should be built, but I'd love to see and play with any experiments in this space. -- [0] - https://www.kialo.com/tour/ https://www.kialo.com/tour/

Folcon 8y ago
Ok that's interesting, you've clearly done a lot in this space already, are you still working in this space and would you be up for a quick discussion about it? I was actually thinking about it from a different direction. Allow people to enter free text and then aggressively parse it out into a structure. Make the structure visible and then work with the structure. But allow the free text to stand as metadata. Allow the person writing to highlight/markup blocks in the main text that support their argument. The structure I was thinking of as a starting point was a list of clauses. You could then surface people stating a clause, saying that it's sometimes true or false, disagreeing with it, and detect it across the nodes in the graph making discovery a bit easier. Or even taking a clause and using it as a starting point for a new discussion. The other thing I think that's necessary is what kind of environment you foster. Something like changemyview[0][1] with the delta system [2] comes to mind. Simply speaking a person awards delta's when they've had their mind changed in any meaningful way by a person, and they're required to give a short explanation or statement about why they awarded it. -- - [0] https://changemyview.net/ https://changemyview.net/ - [1] https://www.reddit.com/r/changemyview/ https://www.reddit.com/r/changemyview/ - [2] https://www.reddit.com/r/changemyview/wiki/deltasystem https://www.reddit.com/r/changemyview/wiki/deltasystem
- TeMPOraL 8y ago
  > Ok that's interesting, you've clearly done a lot in this space already, are you still working in this space and would you be up for a quick discussion about it? Sure! Feel free to mail me (address in the profile). I wouldn't say I have done a lot in this space - I am just interested in it, and try to pay attention to any new ways to structure discussions and thoughts that helps people reach agreement. -- Your concept is very interesting, and indeed a completely different direction than what I was thinking. You seem to want to start from making free-text comments more machine-understandable, so that they can be further analyzed and grouped with similar points from other people? This seems it could mesh with what I wrote in the previous comment - I want to get a graph of ideas, and you just suggested an approach to generating that graph.
  - 8y ago