Comparing versions
Version 6Current version
The moderation package is intended to allow for more user participation in the journalistic aspects of a bitweaver website, ranging from simple user feedback on the quality of a news article to collaborative journalism.

The basics
The core functionality of the moderation package consists of the ability for users to give feedback on the quality of content on a bitweaver website. This is the fundamental function upon which all other functions of the moderation package are based. There are three decisions we will have to make regarding the implantation of this functionality:

1. The types of content we allow to be rated. Primary candidates are articles, comments, wikis, blogs and media files (e.g. images or sound files), but I imagine other types of content might qualify as well (e.g. users). This decision also largely depends on the technical feasibility.
2. The available methods of rating content. There currently appear to be three primary methods of rating content. The most common method is that of simply giving content a rating on a scale of 0-5, often represented by stars. Another method is that of the simple up/down vote, as employed by digg. Finally, a method particularly popular for rating comments is the Slashdot rating system.
The different methods have their strong and weak points, and often websites employ different methods for different types of content (e.g. articles vs comments). We will have to decide which method(s) of rating content we provide and whether we allow administrators choose between different rating methods for different types of content.
3. The final point of concern that affects the basic functionality of rating content, is which factors we allow to influence the rating of content and how they influence this rating. In case of rating a media file, e.g. a photograph, I can imagine that merely taking the average of all user ratings will suffice. Other types of content however, might benefit from having other factors influence the rating as well. For instance, in case of user submitted news articles (e.g. digg), factors like time, frequency of cotes, amount of votes etc would have to be considered as well.
I therefore propose we do not hard-code the formula that determines the rating of content. A better approach, in my opinion, would be to list the possible factors that can be perceived to influence the rating of content and then allow the administrators to weigh these factors based on their unique situation. We could provide default settings per content type as a starting point.

"Additional Functionality"
Based on the aforementioned core functionality, a plethora of very interesting possible additional functionality arises. I will discuss those functions I believe will provide the most additional value for bitweaver, but I would really enjoy others suggesting different functionality that they might regard as even more important for bitweaver.

The first thing I would like to suggest is to utilize user feedback to support the approval process of user submitted news articles. While the currently employed centralized approval model works well for certain types of websites, other types of websites might want to employ a more decentralized, user-centric approach to news articles. I would therefore suggest to expand the current functionality with two additional "optional" news article approval systems based on user feedback (although technical and time constraints could very well force us to add only one): semi-automatic approval and fully automatic approval.
The core of both systems consists of the so called "news-room", which is a page where user submitted news articles are listed for anyone to view.

"Advanced Functionality"
A list of functionality that might be interested to implement long-term.

….

"Work in progress!"

 
The moderation package is intended to allow for more user participation in the journalistic aspects of a bitweaver website, ranging from simple user feedback on the quality of a news article to collaborative journalism.

The basics
The core functionality of the moderation package consists of the ability for users to give feedback on the quality of content on a bitweaver website. This is the fundamental function upon which all other functions of the moderation package are based. There are three decisions we will have to make regarding the implantation of this functionality:
  1. The types of content we allow to be rated. Primary candidates are articles, comments, wikis, blogs and media files (e.g. images or sound files), but I imagine other types of content might qualify as well (e.g. users). This decision also largely depends on the technical feasibility.
  2. The available methods of rating content. There currently appear to be three primary methods of rating content. The most common method is that of simply giving content a rating on a scale of 0-5, often represented by stars. Another method is that of the simple up/down vote, as employed by digg. Finally, a method particularly popular for rating comments is the Slashdot rating system. The different methods have their strong and weak points, and often websites employ different methods for different types of content (e.g. articles vs comments). We will have to decide which method(s) of rating content we provide and whether we allow administrators choose between different rating methods for different types of content.
  3. The final point of concern is which factors we allow to influence the rating of content and how they influence this rating. In case of rating a media file, e.g. a photograph, I can imagine that merely taking the average of all user ratings will suffice. Other types of content however, might benefit from having other factors influence the rating as well. For instance, in case of user submitted news articles (e.g. digg), factors like time, frequency of votes, amount of votes etc would have to be considered as well. I therefore propose we do not hard-code the formula that determines the rating of content. A better approach, in my opinion, would be to list the possible factors that can be perceived to influence the rating of content for every content type and then allow the administrators to weigh these factors. We could provide default settings per content type as a starting point.
Automated Article Approval
Based on the aforementioned core functionality, a plethora of very interesting possible additional functionality arises. I will discuss those functions I believe will provide the most additional value for bitweaver, but I would really enjoy others suggesting different functionality that they might regard as even more important for bitweaver.

The first thing I would like to suggest is to utilize user feedback to support the approval process of user submitted news articles. While the currently employed centralized approval model works well for certain types of websites, other types of websites might want to employ a more decentralized, user-centric approach to news articles. I would therefore suggest to expand the current functionality with two additional optional news article approval systems based on user feedback (although technical and time constraints could very well force us to add only one): semi-automatic approval and fully automatic approval.
The core of both systems consists of the so called "news-room", which is a page where user submitted news articles are listed for anyone to view and rate (example: digg newsroom). The actual rating of an article is not only determined by the average of the user rating feedback, but also (if deemed necessary) by the frequency of the feedback, the amount of feedback and the age of the article. Once an article reaches a certain (administrator defined) threshold rating, it is marked for display on the main news page. At this point the difference between the two systems becomes apparent: in case of the fully automatic approval system, the article will automatically be moved to the front page without require any actions by the administrator. However, in case of the semi-automatic approval, the final choice approval of the article lies with the site administrator/moderator.

Automated Moderation
Another example of what might be very useful, is the ability to automatically moderate content based on ratings. This would be very useful in fighting spam and other types of unwanted user input on a bitweaver website. We could offer two levels of automation: semi-automatic and fully automatic moderation.
In case of semi-automatic moderation, content (e.g. comments) hat receive a low rating a certain number of times, cause the website administrator to be notified (e.g. by email). A seperate page, similar to that utilized by the current articles approval page, could be used to list all content that requires attention from the administrator along with easy delete controls.
Fully automatic moderation would automatically hide content that receives a low rating a certain number of times, and notify the administrator afterwards. The administrator would then have ability to undo the moderation of the content utilizing a similar approach as with semi-automatic moderation.

Filter
Quite a few websites that allow for user input (e.g. slashdot) also allow users to determine the quality of content that is displayed with use of a filter. Bitweaver could allow administrators to enable certain (or all) types of content to be filtered by users. This would be rather useful for bitweaver websites with a lot of user comments or user submitted media content.

Moderators
Given all the additional tasks that will have to be performed by the administrator when all the functionality of the moderation package is enabled, I think it would be a good idea to allow administrators to create a type of user dedicated to moderating the website. This type of user, the moderator, would perform all functions related to making sure that the user input process proceeds smoothly. I belive this type of user would be greatly appreciated by website owners who want to employ not just administrators with rather broad permissions, but also editors/moderators with less permissions (and hence less chance to break something).

Advanced Functionality
In addition to the previously mentioned functionality that could be added, there's also slightly more advanced functionality possible that might be a bit tricky to implement but interesting nonetheless.

The first interesting idea is to give users additional influence based on positively contributing to the website. This could be measured by e.g. the average rating of the users' contributions, the amount of contributions or the amount of moderation performed. Based on their merit users would be given additional weight to their input. Another possibility is to publically display the average rating of a users' contributions to demonstrate the value of the user to the community.

A second interesting idea relates to automatic moderation of wiki pages. This would be made possible if bitweaver would not only keep records of the average rating of a wiki page, but also the average rating for every revision. This would help to detect vandalism, spam or simply low-quality user contributions. If a signifgicant drop in the rating of the wiki would be detected, the wiki could either automatically revert to a previous version or notify a moderator/administrator. In combination with keeping tally of the average quality of a user's contribution, this would also help detect users that consistently input low quality contributions (e.g. spam or vandalism).

The final suggestion I'd like to make would be to combine automated approval of news articles with automated moderation of wiki pages. By keeping news articles editable as if they were wiki pages while they are waiting for approval in the news room (or even after approval, depending on administrator preference), the quality of the article quality could be improved by a distributed effort until it meets the quality standards deemed necessery.

I hope this overview will get a discussion started and I look forward to your ideas and critiques!

Page History
Date/CommentUserIPVersion
17 Feb 2006 (10:19 UTC)
AC van Rijn146.50.202.12614
Current • Source
AC van Rijn146.50.202.12613
View • Compare • Difference • Source
AC van Rijn146.50.202.12612
View • Compare • Difference • Source
AC van Rijn146.50.202.12611
View • Compare • Difference • Source
AC van Rijn146.50.202.12610
View • Compare • Difference • Source
AC van Rijn146.50.202.1266
View • Compare • Difference • Source
AC van Rijn146.50.202.1265
View • Compare • Difference • Source
AC van Rijn146.50.202.1263
View • Compare • Difference • Source
AC van Rijn146.50.202.1262
View • Compare • Difference • Source
AC van Rijn146.50.202.1261
View • Compare • Difference • Source