This document is available on the Internet at: http://urbanmainframe.com/folders/blog/20040622/
Content syndication is a wonderful, time-saving innovation and I am grateful to every single content producer who provides RSS feeds from their website.
The days of having to maintain a vast, unmanageable list of bookmarks (or favourites), and of having to periodically visit hundreds of websites to hunt for changes, are truly over.
Users are able to enjoy a wider, more diverse web when an RSS reader is taking care of the drudgery that searching for new content entails.
However, "every rose has it's thorn" and, when that thorn is a poorly configured RSS feed, then the rose begins to lose its allure...
“the problem lies with the publishers and it is this: insufficient granularity”
I am currently monitoring just over 100 websites via the combined efforts of my blogroll and RSS reader. Now 100 websites is not really that many. Yet, even when tracking this small number of websites, I keep finding myself frustrated with syndication.
There's nothing wrong with RSS per se, nor is there anything wrong with my reader. The main problem lies with the publishers and it is this: insufficient granularity.
In the early 21st century we are all suffering from information overload. Media surrounds us and is pushed at us from all angles, data is fed to us in ever-increasing volumes and our productivity is threatened by the vast amount of information our brains have to process on a daily basis. At any given, waking moment, there are a multitude of information sources all clamouring for our attention.
We manage this confusion with selectivity, compartmentalisation and self-discipline. But we can only exercise the second and if we are able to practice the first and, sadly, some RSS feeds restrict our selectivity.
More than half of the feeds I subscribe to are incorrectly time-stamped. That is, their time-stamps are updated every time the content is edited instead of reflecting the publication date of the content. The RSS 2.0 Specification describes "pubDate" as "indicates when the item was published," not "indicates when the item was last updated".
Don't distract me, don't break my concentration just because you've corrected a typo on your website. Leave the "pubDate" alone after the initial publication. If you must let the world know that you've added a comma to your latest article then, please, do so via a separate "revisions" feed, or keep a changelog.
This is the killer for me personally. This is the one that usually pushes me in the direction of the "unsubscribe" option. I suspect that everyone with an RSS reader is familiar with the problem... feeds that index disparate types of content in one, all-encompassing file.
If you are going to syndicate all of your content (articles, blog entries, links, blogroll, etc), offer separate feeds for each content type - or use descriptive titles to distinguish types.
Consider the following examples and ask yourself which is the most appealing.
Read, reference and comment - an online format that does it all
The world's two worst variable names
Gmail: Desirable or Not?
President Bush Greats France's Jacques Chirac
[Link] Read, reference and comment - an online format that does it all
[Link] The world's two worst variable names
[Weblog] Gmail: Desirable or Not?
[Image] President Bush Greats France's Jacques Chirac
The second example is much more useful to me than the former.
Give your readers the information they need in order to exercise selectivity, or provide the reader with distinct "channels" of content.
Your subscribers will appreciate the choice.