Always New Mistakes

December 13, 2007

Idea: Embedding RSS comments in a content RSS feed

Filed under: Blogs — Tags: , , , , , , , — Alex Barrera @ 4:05 pm

Mark Evans had a nice question on his blog some time ago and while commenting it I got an idea. One of the biggest problems (among others) about blogs, comments and RSS is that you can’t have both at the same time. That is, you can’t download one RSS feed with the content and the comments (as far as I know, so if I’m wrong, please correct me). I am aware that some feed readers can display both, but it’s not mainstream yet. On of those systems is the one on which I had the luck to see at FOWA in London this year. They are doing an awesome job and their interface is very very cool. Anyhow, my idea is based on something I read on the latest OPML draft specification (2.0) by Dave Winer.

Most blogs and content platforms offer two RSS feeds. One for the actual content and another one for the comments. The problem with it is that you need to subscribe to both to get them AND most probably, your feed reader will display them as separate feeds. In my opinion this is a burden and it might be one of the reasons why we seeIdea lightbulb such low ratios of comments per subscriber. I always tell the same fact, but it’s astonishing, at least for me, that a blog like TechCrunch with 165k+ subscribers gets an average of 20-30 comments per post. So, looking at the RSS 2.0 specification I remembered that there is a <comments> tag for each RSS item where, in theory, you have a link to the comments. Now, getting back to the OPML draft, Dave added a new feature which I think could allow a better comment handling. The basic tag in OPML is the <outline>. This tag might have an attribute called type, which can take several values, among them the “link” value. Quoting the OPML specification: “An outline element whose type is link must have a url attribute whose value is an http address“. In the new specification, he added what he calls inclusion. That is, if the url of an outline with link type ends in .opml, the file will be downloaded and included when displayed. I think the idea is very cool, specially if we give it a little twist. I was thinking that if instead of doing the inclusion with .opml files, we do it with an RSS file and we add that capability to the <comments> tag in the RSS specification we could have a wonderful tool to include comments in a feed.

So, putting all this together we might end up with something like this:

An RSS file of a blog ( which has items like the following one:

<title>Twittories and the art of writing</title>
<pubDate>Thu, 13 Dec 2007 14:03:55 +0000</pubDate>

Now, if instead of having a comments tag like that, we mix it with the OPML notion of outline included link we could rewrite it as:

<comments number="2"></comments>

With that slight modification we can now have the number of comments we have and the xmlURL of the comment’s RSS feed for that post. What happens on the other side? Well, the feed reader must read that tag and include it as if it was part of the feed. Adding the number attribute allow us, on one side to show the readers how many comments the post has and on another hand lets the feed reader know if he needs to download the RSS feed of the comments again. Evidently, the number attribute will change with time if we have new comments and will be updated every time our feed reader requests the RSS feed.

So, what do you think of my idea? As with every thing I’m sure this issue has been discussed by someone before me and it might even have a different solution I’m unaware of. So please, if you know a cool way to handle comments drop me a comment and I’ll include it in the post.

UPDATE: This is what happens when you rush a post. You get it wrong. What I just described exists already and is known as wfw:commentRss. The only difference with what I was proposing is that the comment count is specified by a different tag (slash:comments), which by the way has nothing to do with the wfw namespace. Now, the commentRss tag was proposed in 2003. I’m quite fascinated as to why it hasn’t been adopted in feed readers yet. Some readers display the comments count, some others even parse the commentRss tag, but I don’t know yet of a reader that actually pulls that data based on the counter and collapse it with the content. Maybe you could even ask the user to activate that feature on a per post basis.

Twittories and the art of writing

Filed under: Business — Tags: , , , , — Alex Barrera @ 3:03 pm

Some days ago, our good friend Duncan Riley wrote about Twittories. It’s a website where many people collaborate to write a short story. The catch is that they have to do it with a twitter message. This means, each user can only submit one message with a maximum length of 140 characters. While reading the story I was reminded of a game we used to play when we were little, I know it has another name in English but can’t remember now. In Spanish it was called the broken phone. Anyway, although it’s a fun idea and a cool experiment, there is no way you can read a story written in that fashion. The central plot is so broken it’s impossible to follow. But I must say it got me thinking about the quality of most articles. Before creating my current startup, I developed a prototype for what was going to be my first try in the entrepreneur world. I designed a semantic algorithm that was able to detect and highlight structural problems in texts. In other words, a style corrector. While I was testing it, I realized how badly written where most news articles. One of the most common problem was the use of overly large phrases in terms of words. And when I talk about large phrases I’m talking about 50 or more words per phrase without a full stop. So, back to the twittories, I realized that, even though the plot might be messed up, the 140 characters restriction could be a great thumb rule for many writers (including myself of course). Maybe we could enforce some rules like that on our blog platforms. Now that TypePad is also open source, there is no excuse 😉

What are your thoughts about this?

Blog at