Info Architecture
Weblogs should be topic-first, not author-first
Clay Shirky (via Ross Mayfield):
"The weblog world has taken the 4 elements of organization from mailing lists and usenet -- overall topic, time of post, post title, author -- and rearranged them in order of importance as author, time, and title, dispensing with topics altogether."
This is something that makes me a little queezy about weblogs. If I had my way, I'd rank the importance of topics as number 1. I would like my RSS Aggregator to deliver RSS feeds to me based on topics that I subscribe to, rather than by author. Don't get me wrong, an author's 'voice' is what makes a weblog unique and interesting. But I don't believe the blogosphere should rank the importance of a weblog post based on who the author is. Google ranks on topic relevance, why shouldn't the blogosphere?
For example, I am very interested in reading weblog posts about Longhorn. But I actually don't care who writes them. I'm curious about everyone's opinion - and I will make up my own mind about how relevant the information is. Right now I rely on Robert Scoble's weblog to deliver me new posts about Longhorn. Robert Scoble has a fantastic policy of giving everyone "one free link" and I enjoy the wide variety of weblogs he links to. It gives his weblog a feeling of openness and freedom. Everybody - A, B and C List bloggers alike - feels welcome to respond in writing via his Comments. It's no coincidence that Scoble's comments system is one of the liveliest and interesting on the Web (even though it is a bit funky - ie the software deletes stuff every now and then!).
Clay Shirky goes on to say:
"This "author-first" organization gives the weblog world a huge boost, as the "Who said what" reputation system we all carry around in our head is a fantastic tool for organizing what we read, as well as acting as a kind of latent bozo filter."
I couldn't disagree more with that statement - organizing what we read according to who wrote it is plain elitist. I'd much rather organize what I read based on topics - then decide for myself if it has any value. This is the beauty of a system such as k-collector, which Paulo and Matt have developed as a means to track and connect peoples weblog posts by topic. It's still in development, but so far I like what I see. I can browse topic pages and read through what various bloggers have to say on the same topic. It's a good way of discovering new voices - rather than simply reading the A List Bloggers.
To be fair, Clay does point out the one major downside of the "author-first" model of weblogs - the blogosphere becomes riven with personality clashes. This is particularly so amongst the A List Bloggers, who because of their high profile in the blogosphere have reputations to uphold and so frequently challenge each other. They engage in epic pissing contests and unfortunately the rest of us get sprayed too. It's not particularly fun to read and, as Rogers can testify, it's no fun to be on the receiving end.
Watching the Echo project unfold in the blogosphere is like watching the WWF's Royal Rumble - we see each A List Blogger climb into the ring, until about 30 of them are in there slugging it out. Who will be left in the ring at the end? Who knows, and I don't care. I'd rather just read and write about topics that are of interest to me, thanks. Leave the Web-wrestling to the pros (it's all fake anyway).
How can we improve the weblog world, so that it does take into account topics? Well I'd like to see new versions of RSS Aggregators give us some more options to filter weblog posts by topic. And Paulo and Matt, I look forward to the next version of k-collector. Keep up the good work!
Building on Google's aggregation services
In his article "Google Aggregation Strategy", Elwyn Jenkins from Microdot News reviews three Google "information aggregations" and asks which one will be moved from beta to live first - Blogger, Froogle or Google News. Microdot News argues that aggregation is at the heart of Google's business and that Google will build on the success of their search engine by offering similar aggregation services for weblogs, shopping and news - and more.
I will take that argument one step further, because I think that Google will start to provide syndication services as well as aggregation. The current crop of Google's aggregation services are "pull" rather than "push". People still need to go to Google's websites in order to find information. I suggest that Google will adopt a publish-subscribe model. People will subscribe to information and it will be automatically delivered to them on a regular basis.
The aggregation of information is the base for Google's success. But the value-add is to enable people to tap into Google's aggregated information base and create personalised "feeds". These feeds won't be the same as the results from traditional search queries. Instead of inputting a wide-ranging query, people will enter a somewhat more constrained "topic". For example rather than entering an esoteric query like "Dave Winer XML-RPC spicy noodles in Boston RSS sunset", a person will need to enter something that can be converted into a reasonably generic topic. A "topic" in this sense will be something between a Google search query and a Yahoo category. That is, it will be automatically generated but within a manually-defined framework (such as ENT perhaps).
RSS - Subscribing to Topics
RSS stands for Really Simple Syndication, or Rich Site Summary, depending on who you listen to. Either way RSS has become the poster child for the Publish-Subscribe protocol. RSS allows weblog and website owners to syndicate their content to anyone who wishes to subscribe to it. Usually people subscribe to "RSS feeds" via an aggregator. Some weblog authoring tools also support RSS subscriptions - for example Radio Userland.
The weblog community has embraced RSS. Most bloggers have RSS feeds now - even design guru Jeffrey Zeldman has signed up (and immediately defended himself against accusations that he had "sold out"!). With RSS there is less control over design - it is enough work designing for different browsers, without having to cater to multiple aggregators too! Text is primary with RSS, which I believe brings ideas to the fore.
RSS currently is used to syndicate personal content or news items. But there is another value in RSS, which is beginning to be explored more - using RSS as a means to subscribe to content by topic. RSS topic subscriptions bring us closer to the Semantic Web - defined by Tim Berners-Lee as "an extension of the current Web in which information is given well-defined meaning". Subscribing to an RSS topic feed lets a web user accurately and automatically gather information that has specific meaning to them.
Most RSS feeds currently are personal weblog and news items - which means the publisher controls the content. RSS feeds that are based on topics on the other hand, give power back to the subscriber. As an example, recently I subscribed to the BBC top stories RSS feed. But there are only a few topics I'm interested in from BCC - news on the Iraq war, British football, any technology stories that come up. I found though that a lot of the content was local British news, which I have little use for since I come from New Zealand. Eventually I unsubscribed from the BBC news feed, because there was too much content that was irrelevant to me. But wouldn't it be great if I could subscribe to only the news items from the BBC that matched the topics I specified? I could enter "Premier league football" as a topic, or I could even narrow it down to my favourite team "Manchester United".
I would like to be able to subscribe to information that currently interests me. For example I am interested in the "two-way web". I can subscribe to personal weblogs that I know will discuss that topic from time to time. But how much more efficient to simply enter the phrase "two-way web" into my RSS aggregator and have it scour the Web, finding and delivering articles about the two-way web. It would be like entering a search into Google - but because this is RSS it is an ongoing, automated delivery of information to me.
There is also an increased chance that I will find more gold nugget articles, by writers I've never heard of. RSS topic feeds may go some way to addressing the A-List blogger problem, whereby a small percentage of bloggers get a disproportionately large amount of hits. This is due to what's called a power law in network theory - a few hubs in a network are immensely more popular than the majority of hubs. Power laws in fact define many things such as the movie industry and the music scene. The World Wide Web itself is a prime example of a power law in motion. A tiny number of "hubs" like Google and Yahoo get a vast number of visits every day. But the Web is made up of millions of pages, most of which are hardly ever visited. (ps check out this Google search query for an ironic example of power laws on the Web - webloggers are the top 3 search results!).
RSS topic feeds will still more likely pick up items from popular bloggers, but at least there is a chance the rest of us will have an article or two "subscribed" to via a topic request.
I want to finish by mentioning a couple of interesting projects that are moving us closer towards a topic subscription model for RSS.
One such project is ENT - which stands for "Easy News Topics". ENT is a module of RSS2.0 and the goal is to "enable a new generation of aggregator applications to be written. These aggregators will allow people to filter items they do not wish to see, prioritize those about things they are interested in and recombine items into new feeds." ENT relies on RSS publishers entering new tag data, in particular the tags 'cloud' (a URI source which describes the topics) and 'topic' (a "metadata item"). There are issues with this "self-categorization" - reliability of data, differences in how people interpret words and phrases, etc. However one of the ENT authors Matt Mowers responded by saying that "ENT could just as easily be used by a categorizer bot that sucked in feeds and annotated them (using heuristics) with topics from it's own cloud." This is an intriguing idea - an automated Yahoo of the RSS world?
Another new RSS app that gathers information by topic instead of author, is Gnews2RSS at VoidStar.com. This application creates RSS feeds on-the-fly for Google news items. You simply enter a topic into a search box, then click the "Create RSS" button. An instant RSS feed is displayed, allowing you to subscribe to it on an ongoing basis using your favourite RSS aggregator.
RSS subscriptions by topics is a step towards the Semantic Web. It gives more customization control to web users and allows all writers to reach a wider audience.
