For a long time, I’ve been thinking superficially about the interaction design of commenting systems on the web. I’ve long been irritated by the structure of basic commenting tools and am sure most other web folk feel the same way. The problem is so bad that currently, there are rarely any articles with 20+ comments which provide a useful addendum to the main article. One assumption to note here is that my irritation arises in the context that a commenting tool is a means to discuss, debate and add to the main article.
In the current paradigm for commenting, there is a linear list of comments – a new comment is left directly below the previous comment. The primary metaphor here is a linear sequence – a line.
As I probed further into what irritated me, I could list out several problems:
1. Missing context:
There is no indication of who a new comment is for –
Is it for the article?
Is it a simple expression of appreciation?
Is it debating a point addressed by another commentor?
Of the several debates happening on the article, which one is it catering to?
This causes confusion when there are over 10 comments because a new commentor would either have to read through all the comments, many of which will not be of interest to her, and decide which one to comment on.
Or she will create an entirely new point of discussion further adding to the chaos for the next commentor.
2. Quality of discussions:
This is related to the point above – the linear stream of comments is interspersed with disparate points of view – some tiny comments of appreciation, some discussing one interpretation of the article while others going off on new tangents.
Over time, this stream of comments completely loses its value thanks to linear list of disparate ideas.
3. It’s incredibly difficult for a discussion to evolve around an article.
Take a simple example – I’ve written an article supporting the point that the color blue is better than the color red. Now the first commentor agrees with my position and adds to it.
The second commentor takes issue with it and attacks it.
The third user is irked by a point made by the first user and writes about it.
A fourth commentor appreciates my style of writing and expresses that.
A fifth commentor adds to the second ones argument and attacks the first one.
Now, imagine this going on for another 20 comments. And then imagine what happens when the first user comes back to this stream to see how the discussion has evolved – he will see a stream of incomplete threads and would have to filter them to find out the ones he really cares about. Oh what pain!
The chances of him abandoning his pursuit at this point is extremely high.
There are several such synonymous problems with current commenting systems.
There have been efforts to tackle this such as:
1. A dedicated Like/Dislike buton to keep the tiny appreciative/disparaging posts at bay.
2. A nesting system to allow commentors to reply to specific comments rather than disrupting the entire flow of other discussions.
3. An upvoting/downvoting system coupled with nesting to bring the best discussions on top of the list.
(hackernews, stack overflow)
However, there hasn’t been a radical shift in the way commenting systems are implemented.
It’s time for a new paradigm.