With December just around the corner, you can bet your bottom dollar (or even just your bottom) that I’m in the throes of preparing for yet another year of 24ways. Just as we’ve done since 2005, we’ll be serving up a new web design or development article each day for the first 24 days of December. Consider it something like a geek advent calendar, except without the little doors.
This year, we’re aiming to try something a little different with the comments. Comments on articles can be great, but sometimes they can add more noise than value. This is especially true if an article gains traction and is linked to in places like Digg or Slashdot, bringing in fresh readers without any context. Whereas in many places it’s not uncommon to see hundreds of comments per post, we’d really like to encourage an environment of well-considered commentary that actually adds something to the conversation.
So how do we go about encouraging that? I think the first step is to make people take responsibility for their words, just as the author of each article does. Comments can be anonymous, and often the commenter doesn’t have a good sense of who might be reading, so the first thing we’re thinking of doing is making the commenter’s own site or weblog the primary place to post a comment. Each article will carry a tag, and 24ways will aggregate posts from the web (via Technorati) based on that tag. Sort of like a trackback, but less revolting. More like Tagbacks. Ew.
Of course, not everyone who has something worthwhile to add has their own site where they can post and tag, so I think we’ll need to keep comments open. However, we’re thinking of reducing the weight of the comments, and moderating so that by default nothing gets published. If a comment is well considered, it’ll be published. We’d be adding the signal, not removing the noise.
There’s a few other ideas we have brewing, but I’d really appreciate any feedback on this. As always, it’s a fine line between making sure comments are adding to the conversation, and discouraging people from commenting at all.
Make sure you’re subscribed to the RSS feed ready for December. I think it’s going to be fun.
Comments
I think that the fact you have a daily release schedule for the project might lend itself to a daily round up of all the comments. It’s more work for you of course, but it means that you don’t have 50 people saying the the article is cool just so they can get a link back. Also, if you use “no follow” on your comment links, let people know I’m sure that’ll help cut down on the dross. Looking forward to this year’s calendar.
Sounds like a good way to manage article discussion… straight commenting can become a mess of lost context. Encouraging people to comment with blog posts will certainly reduce the “me too” posts.
This sounds like an excellent idea. I have something similar on my blog posts but you have to go through an extra step (clicking on “Who’s linking to this”). And I really like the idea of editorially controlling comments.
Will you be providing a machine tag for the “tagbacks”? That’s what I’ve done for the Flickr integration over on adactio.
Oh, and should I be offended that I haven’t been asked back to write something this year? ;-)
I hadn’t really thought about whether it should be a machine tag or a regular tag. I guess the tricky bit it that I need to encompass three items; the namespace (24ways), the year and the day. Perhaps that could be:
24ways:article=2007-01
for the first article of 2007. However, we also have the fact that articles map neatly to dates, so perhaps we should use the date:
24ways:date=20071201
Or something else. This clearly needs some thought.
I hope you’re not offended, Jeremy! I’m attempting to keep the line-up fresh by mixing some new names with familiar names. You’re not outcast, you’ve just been left to fallow :-D
Why invent a unique machine tag for each thing? Each post already has a unique identifier – its URL – which already exists, is trackable through technorati, and doesn’t require everyone to ignore how the web works and reinvent linking in order to use it :-)
Must actually write mine…clearly Drew loves me more than Jezza, despite how I’ve got worse hair…
Stuart – there’s a big difference between linking to something and making a comment about it. People frequently link to webdev articles from their weblogs, sideblogs, tumbleblogs, booking marking app and all sorts without adding anything to the conversation, and that’s great. Love it. Thanks for the link! But there’s no value in listing that and linking back to it, as it’s not a response or comment. That’s where trackbacks fall down (one place at least).
Tagging gives an explicit way to label a post as a response.
I think you’re going to be able to shape your responses in this way. Of course anonymous comments are needed and valid on many sites. But if you want real, substantive input about your posts, it’s completely up to you if you wish people to post their real names. Tagging back to their own pages will just be another level of legitimizing that person.
Drew: that’s what is for—there was discussion about this at @media (I think it was at the speakers dinner, though). I reckon we seriously need to think about how best to solve this problem: sticking a machine tag in and saying to people “mention this string in your response” is low-barrier-to-entry, but it strikes me as not really being in tune with the karmic ebb and flow of how the web works :(
Stuart – I think my comments system might have eaten part of your response. That’s what what is for?
I’m all for finding the best solution – that’s why I’ve asked for feedback :)
I was only thinking about this year’s 24ways yesterday. Good to hear its happening again. Can’t wait for some informative posts to lead me up to Christmas.
Ah. That’s what something like:
[a rel=”comment” href=”http://24ways.com/2007/1”]
or similar is for, I think. Machine tags (and asking people to mention them so they’re searchable) seem like a way of doing linking that doesn’t involve links…
Whilst that’s true (although I’m pretty sure it should be rev=”comment”) in practise I’m not sure it’d work. I don’t have hard facts, but I imagine that most people aren’t blogging by hand, but are using some kind of publishing tool that writes HTML for them. I’d be dead impressed if any such tool supported the adding of the rev attribute on links in any sort of convenient way.
Many common blogging tools, in contrast, do support tagging in a user-friendly way. Plus there are useful web services with well behaved RESTful APIs that find those tags for me.
So whilst using rel=”comment” is a good way to show the relationship of a comment to the originating article (and probably the most correct way), I’m not sure that’s going to work in December 2007. And it’s December 2007 we’re talking about here.
Good to hear that 24ways will be happening again – I am looking forward to the daily articles.
You raise a good point about the “noise” of comments that don’t really contribute to the discussion, and I agree. They can be a bother to sift through, especially when there are a lot of them (and there usually are). Your idea of “tagbacks” is an interesting one, and could be effective to gather up posts on the web responding to the original article. And it would allow/encourage commenters to write more complete responses.
Although something that sticks out at me, is that comments are usually interactive – commenters interact with the author of the article, AND they interact with each other. I’m not quite sure if your idea takes that into account. If I understand it correctly, your idea would preclude commenters, on their own sites, from discussing the article with each other, in the way that comments are normally done.
Drew: I understand the pragmatic argument, but…it’s never going to get any better if we just route around the hard stuff rather than fixing it…
I thought the internet was all about routing around problems? :)
Ok, so if I was to use the Technorati cosmos API (the one that finds instances of links to specific URLs) to get pages that link, I could then individually spider each of those pages to find that link and look for a rev attribute. That’s possible.
Do you think that commenters would bother though?
I…am not sure. I hope they would. The 24ways readers skew pretty heavily towards the technical. Way 1 could even explain how to do it, with a short guide to why this sort of microformat matters?
+1 for the rev / rel comments things
:D
So I think what we’ve decided to do is use the intersection of two tags ‘24ways07’ and the author’s name.
The big snag we’ve hit is that the results coming back from Technorati’s API seem to differ wildly from those on their site. Might have to take a different approach.
Interesting solution. I think with the growth of websites like digg, people will increasingly look for different ways to handle comments. Noone wants to read an article where the comments are all things like “first!” “second!” etc. It’s important that readers can continue reading and contribute meaningfully to discussion in the comments.