March 17, 2008 at 01:39Not one karma
Social networks, web 2.0, collective blogs ... Are you startled? I understand you, but nevertheless I invite you to participate in a small brainstorming session. This time I suggest talking about karma / ranking, etc.
So, let's imagine that we have a certain community that unites users within a certain topic and allows us to write to collective blogs.
We need to implement mechanisms that allow this community to independently filter content and decide what is good and what is bad.
It would seem that there is nothing easier. The mechanism working on the same hub All articles are added to the site, there is the possibility of voting. The main article falls after gaining a certain number of points. The mechanism has proven itself remarkably, but there is one big BUT. Be that as it may, the hubr is a geek community on which many sit for hours and it is thanks to them that the articles from the “new” section get on the main page and become available to others.
In our case, such an audience will most likely not be. People come to the site to read the latest news and leave just as quickly. Many just read RSS. That is, most likely, there will be practically no people who will display news on the main page and they will remain unnoticed. In addition, the specificity is such that making people wait until there is an opportunity to write something is also not very correct.
We come to the conclusion that, most likely, the Habr mechanism is not for us.
Nevertheless, we need to somehow protect ourselves from the trash that random passers-by can leave.
Good old pre-moderation is also not our method. This is very annoying when you write an article, add it to the site, and then wait until the moderator wakes up and decides to look at the site.
As a result of such reflections, sketches of his method appeared. I proceeded from the fact that, firstly, our audience is supposed to be quite mature and reasonable, as well as the fact that a person will not always vote for a good article, but with a high probability it will put a minus of obviously garbage.
Instead of “karma”, for each user the concept of “coefficient of trust” (KD) is introduced. For now, let's call it that.
CD is a value that determines the “adequacy” of the author.
Confidence coefficient, unlike karma, has a lower and upper bar. the lower bar is 0 (the author is absolutely inadequate). The upper bar, for example, is 100.
By default, new users receive a certain average value. That is, we still have no reason not to trust them, but it is too early to absolutely trust.
An article whose author’s CD is not less than the allowable border (the allowable border is approximately equal to the new user's CD) automatically gets to the main one.
At the time of writing, the article CD is set equal to the user CD. As a result of voting for an article, its CD changes in one direction or another, taking into account the CD of voters (for example, two negative votes of people with CD 30 and 40 are completely canceled by one positive vote of a person with CD 70).
In most cases, the article will simply go to the main page and will be read / commented on. In rare cases, the article will be minuscule and disappear from the main. Accordingly, the author’s CD will fall and the next time his articles will either fall into the “new” category, or, when the zero CD is reached, he will no longer be able to write to collective blogs.
How to deal with personal blogs is not yet clear. Such a mechanism has only recently come to mind and has not yet had time to mature, but nevertheless I would like to hear your opinions. It may be possible to come up with something new or improve what is.
So, let's imagine that we have a certain community that unites users within a certain topic and allows us to write to collective blogs.
We need to implement mechanisms that allow this community to independently filter content and decide what is good and what is bad.
It would seem that there is nothing easier. The mechanism working on the same hub All articles are added to the site, there is the possibility of voting. The main article falls after gaining a certain number of points. The mechanism has proven itself remarkably, but there is one big BUT. Be that as it may, the hubr is a geek community on which many sit for hours and it is thanks to them that the articles from the “new” section get on the main page and become available to others.
In our case, such an audience will most likely not be. People come to the site to read the latest news and leave just as quickly. Many just read RSS. That is, most likely, there will be practically no people who will display news on the main page and they will remain unnoticed. In addition, the specificity is such that making people wait until there is an opportunity to write something is also not very correct.
We come to the conclusion that, most likely, the Habr mechanism is not for us.
Nevertheless, we need to somehow protect ourselves from the trash that random passers-by can leave.
Good old pre-moderation is also not our method. This is very annoying when you write an article, add it to the site, and then wait until the moderator wakes up and decides to look at the site.
As a result of such reflections, sketches of his method appeared. I proceeded from the fact that, firstly, our audience is supposed to be quite mature and reasonable, as well as the fact that a person will not always vote for a good article, but with a high probability it will put a minus of obviously garbage.
Instead of “karma”, for each user the concept of “coefficient of trust” (KD) is introduced. For now, let's call it that.
CD is a value that determines the “adequacy” of the author.
Confidence coefficient, unlike karma, has a lower and upper bar. the lower bar is 0 (the author is absolutely inadequate). The upper bar, for example, is 100.
By default, new users receive a certain average value. That is, we still have no reason not to trust them, but it is too early to absolutely trust.
An article whose author’s CD is not less than the allowable border (the allowable border is approximately equal to the new user's CD) automatically gets to the main one.
At the time of writing, the article CD is set equal to the user CD. As a result of voting for an article, its CD changes in one direction or another, taking into account the CD of voters (for example, two negative votes of people with CD 30 and 40 are completely canceled by one positive vote of a person with CD 70).
In most cases, the article will simply go to the main page and will be read / commented on. In rare cases, the article will be minuscule and disappear from the main. Accordingly, the author’s CD will fall and the next time his articles will either fall into the “new” category, or, when the zero CD is reached, he will no longer be able to write to collective blogs.
How to deal with personal blogs is not yet clear. Such a mechanism has only recently come to mind and has not yet had time to mature, but nevertheless I would like to hear your opinions. It may be possible to come up with something new or improve what is.