Thread View

The current discussion group thread view mechanisim is less than ideal at conveying the maximum amount of information in a minimum amount of space. Using minimal space is essential if the link macros are to be used in interesting locations, such as in a right hand column. Cramming lots of information into a small area requires care and attention to exactly what information is important. I’m trying to model my interactions on Edward Tufte’s information presentation theory.

The first example below (unimplemented) each asterisk would be a link save the current page. The second and third examples are implemented and can be downloaded at the respective links. The thread listed in the examples is the current thread.

One Idea: “Thread View #1, using UL” “Thread View #2, using pre”.
(Not Implemented)

{threadView()} {threadView2()}

The principles that I’m going on are

  1. No redundant information
  2. Compact, so we can show as much relationship information as possible
  3. Reduce marginally relevant information
  4. Bring important information to the foreground.
  5. Allow additional information with user interaction.
  6. Every pixel conveys information by position, color, and content.


Titles are not relevant information. Most of the time, they are the same as the message that is being viewed. Perhaps differences in the titles should be highlighted when the changes are more significant than a missing Re: or a change in capitalization. (principles 1,2,3)

The relationship to the whole thread is important. Visually identifying the current position within the thread by not linking the current document (putting it in a single bold line?) is essential. Distinct colors for unvisited and visited links is also essential. (3,6)

To a reader familiar with the community, the author is typically the best predictor of the value of an article. Using only the first name of the author can convey a significant amount of the possible information within the smallest possible space. Even where there are multiple members with the same first name in a community, the consequences of confusion are small. (4,5)

Revisiting the first model, I’m not convinced that * is a good link anchor. The character itself does not convey any useful information about the content or revelancy of the article. In place of the author name, there could be some other small piece of relevant data. Perhaps a count of page views of the message is a valid measure of the usefulness of the article.

No comments

No comments yet. Be the first.

Leave a reply

You must be logged in to post a comment.