Ask HN: How to design a karma reward system • Hacker News Board

Re: Ask HN: How to design a karma reward system

#21

Post by yannis » Mon, Dec 28, 2009, 5:41 AM UTC

Perceval wrote: post #10

I've posted this on HN before, but I wrote a long essay about problems with account-karma-moderation systems: http://www.kuro5hin.org/story/2009/3/12/33338/3000 Probably too involved for your project, but there are some problems you should be aware of from the get go.

Perceval thanks and by the way an excellent essay. You hit the nail on the head with ... 'our hypothetical comment-driven forum would push stories up based on quality of conversation....'. Do you think that part of the formula can be improved by including automatically a 'readability formula', where the childish comments and language would automatically be penalized by the system?

Re: Ask HN: How to design a karma reward system

#22

Post by Perceval » Mon, Dec 28, 2009, 6:16 AM UTC

yannis wrote: post #21

Perceval wrote: post #10

I've posted this on HN before, but I wrote a long essay about problems with account-karma-moderation systems: http://www.kuro5hin.org/story/2009/3/12/33338/3000 Probably too involved for your project, but there are some problems you should be aware of from the get go.

Perceval thanks and by the way an excellent essay. You hit the nail on the head with ... 'our hypothetical comment-driven forum would push stories up based on quality of conversation....'. Do you think that part of the formula can be improved by including automatically a 'readability formula', where the childish comments and language would automatically be penalized by the system?

Some folks are already working on that: http://stupidfilter.org/main/

It's supposed to identify formally stupid comments, i.e. those that have bad spelling and txt abbreviations, etc. It doesn't do anything about substantive stupidity, unfortunately ;)

Re: Ask HN: How to design a karma reward system

#23

Post by yannis » Mon, Dec 28, 2009, 9:32 AM UTC

Perceval wrote: post #22

yannis wrote: post #21

Earlier quoted context omitted.

Perceval thanks and by the way an excellent essay. You hit the nail on the head with ... 'our hypothetical comment-driven forum would push stories up based on quality of conversation....'. Do you think that part of the formula can be improved by including automatically a 'readability formula', where the childish comments and language would automatically be penalized by the system?

Some folks are already working on that: http://stupidfilter.org/main/ It's supposed to identify formally stupid comments, i.e. those that have bad spelling and txt abbreviations, etc. It doesn't do anything about substantive stupidity, unfortunately ;)

I was thinking something like the Gunning fog index http://en.wikipedia.org/wiki/Gunning_fog_index