Growing Pains and Disqus

Disqus is definitely one of the most powerful commenting systems out there.  Beyond having fabulous customer service, it also is seeming to promote a really social commenting community on blogs, as seen from Fred Wilson’s (currently) 183 comments here.  Whatever solution to solve the overwhelming feeling of seeing that many comments has to include dynamic viewing, to reward good comments over bad ones, and to keep comments rolling.  Further considering that Disqus is a social commenting system, rather than just an ordinary commenting system, if one were just to compress the comments in some sort of ordinary fashion, there runs a risk of losing what makes them inherently valuable- the conversation.

Any system developed has to emphasise where conversations are going, and hide where the conversation has died down.  In essence, Disqus has to develop ways at this point to become an even more dynamic system, even though it is already the most dynamic blog commenting system to date.   This dynamism has to be done in order to  hide the unnecessary from regular readers, and give good introductions to those who just arrived.

For really popular blogs running on the Disqus system, this actually is a problem.  Who wants to sit there and read all 183 comments, as is occurring on Fred’s blogs?  What if you are returning?  How are you supposed to know what comments you’ve read already?  How are you supposed to identify important comments worthy of  returning to?

It is worth taking a look at other systems of commenting on the web, particularly message boards that are very active, such as Slashdot. Copying what works and junking what doesn’t is always a smart idea.  One of the oldest and long lasting ideas for moderation are Karma votes/points.

Although Karma points are in place through Disqus, it isn’t particularly clear how they are enforced.  Are they blog wide, Disqus wide,post wide, or some other way? I recognize that people are recommended when I see profiles, but how to does that link to an individual post on a blog, and how does link to the likelyhood of a comment floating to the top.  Clarifying this system and making this one point clear, and involving Karma points in the blog to keep highly engaged sub-threads on the top of the page might be a good choice.

Currently there is also no system for negative Karma.  If Disqus decides to allow for negative Karma to force a post down to the bottom of the page, it will keep dynamically good comments/and or active comments at the top. This keeps good material and good conversations dynamic, allowing for floating conversations, rather than leaving a page relatively static.

Secondly, since Disqus does cookie users, collapse posts from previous visits except for the first 140 characters down, Ala slashdot.  If someone wants to open them because they are curious, by all means, but they should not be the center focus for someone who is clearly returning back to an old post.  Let the new material take center stage, especially in a long threaded conversation. This will also make it appear that the new material is closer to the top, even if in actuality it is not.

Thirdly break down the comment pages into multiple pages once they hit an arbitrary number of comments and perhaps nudge owners of popular blogs into formats that shrink the text size.  Alternatively, set up that copying the styles-sheet is not an exact copy at the maximum size, but that the comments are one em/pt less in size.  Hopefully with 10 pt/em fonts, the subtle difference will at least make the comment section clearer to the viewer and also compress them down so that they seem less overwhelming

Fourthly, considering that Disqus allows for each comment to have its own unique URL, consider working with someone to allow for semantically tagging and bookmarking the comments when they sidetrack away from the conversation at hand. This will allow even compressed comments to be searchable and boomarkarble, even in a compressed form.  Just because the conversation moves away from some main point, does not mean that the segment is not particuarly valuable, and it should be saved and classified as such.

Fifthly, consider making some features a slow reveal process and a process of surprising the user to keep the cleanliness of the interface.  The more features one adds, the more likely the interface will feel unusable.  Let users develop the skills needed to comment on various levels and to engage on various level of Internet and social network technology.  Not every fit is right for everyone and there is no need to overwhelm someone at first glance.  Give your users a  slow but firm larn curve and they will be alright.

And finally, Congratulations to Daniel Ha and his Team, You guys are amazing for making to the point to create growing pains for others.  You definitely deserve new Poker Chips and Great Cookies.  (perhaps bet the cookies?)

PS If I come up with anything more, it will be posted in a Part 2.

I lied, Two ideas came up:

One that Karma should be a sliding scale courtesy of @ranjeesh

Two is that currently it is functionally impossible in the current system to add good or bad karama to an entire exchange.  Especially when the exchanges are pithy, this is detrimental to the idea of adding up Karma and saving conversations.  Conversations are units just as much as an individual unit of comment is a unit.  A severe annoyance from the perspective of Disqus, because it makes coding around these units all the much harder.  *sigh*  Ideally, a unit of conversation beyond a post should be saved, but I wouldn’t be too upset if that is the limit of code, because how is a computer supposed to know who is talking to who?

Related articles by Zemanta

• Disqus vs IntenseDebate (spicycauldron.com)
• Reactions Now Available In The Comments (avc.com)