If you believe the hype, RSS (including Atom) feeds are the greatest thing since sliced bread. I think they are great, but they still have a lot to learn from sliced bread! RSS still needs the same sort of benefits of increased usability and standardisation that sliced bread brought to bread-making and toasting.<\/p>\n
The following rant is just me carping on about various aspects of RSS (including Atom), WordPress<\/a> and Bloglines<\/a> (my choice in feed aggregators).<\/p>\n Remember: I savagely attack them because I care<\/em>…<\/p>\n WordPress supports a number of different feeds – really quite a surprising number.<\/p>\n The feeds that you can subscribe to include:<\/p>\n WordPress also supports the common feed formats: RSS 0.92, RSS 1.0 and Atom 0.3.<\/p>\n I haven’t (yet?) found the feed I actually want – all of the articles and<\/em> all of the comments.<\/p>\n It is one thing for a web-site to offer a feed. It is another to find<\/em> that feed. How does a page link to a corresponding feed?<\/p>\n The obvious way is to include a hypertext link, visible on the web-site.<\/p>\n The tradition seems to be to use an orange rectangular icon with the letters “XML” in it.<\/p>\n I am having trouble describing exactly how wrong! wrong! wrong!<\/em> that system is.<\/p>\n It is not just that orange so carefully clashes with everyone’s colour-scheme. It is not just that orange appears so prominently, when the icon is only used occasionally by a few people.<\/p>\n It is more that XML is a technical name for low-level file-format. So not only doesn’t it mean the same thing as “subscribe” to non-technical people, it also doesn’t mean the same thing as “subscribe” to technical<\/em> people. You may as well call the link “ASCII” or even “Big-Endian”.<\/p>\n I applaud Microsoft’s decision to ignore the XML icon<\/a> in IE7.<\/p>\n If there is a link with an icon, naive users will click on it to see what it does.<\/p>\n If you click on a link direct to the RSS content, you get a cryptic XML file displayed. (IE6 won’t even open an Atom feed – you need to save it to disk, and open it as an XML file, but that’s another story.)<\/p>\n Cryptic XML displayed to the user? Hold on. Wasn’t that what XSLT was supposed to prevent? Hey, WordPress, where’s the link to an XSLT definition to display the RSS more meaningfully? (Even a static message, explaining to the user where they are, would be a great start.)<\/p>\n Actually, to be fair, WordPress has a different mechanism. By default, it links to the RSS feed like so: The page you are looking for is currently unavailable. The Web site might be experiencing technical difficulties, or you may need to adjust your browser settings.<\/p><\/blockquote>\n Bad IE! Bad, bad IE! You should follow standards…. oh wait up. Maybe for once it isn’t IE’s fault? So, human-visible links to feeds are problematic. XSLT could reduce the impact of the problem, as could the use of more appropriate icons.<\/p>\n If only there was another way – some way the software could automatically detect RSS feeds. Good news! They can, through the use of Note: This is a different type of “link” tag to the typical Where did I find this information about auto-discovery?<\/p>\n Through the help files of the aggregators, like Bloglines? No.<\/p>\n Directly from the standard document published by an authoritative source? No.<\/p>\n From a few random blogs (e.g. Dive Into Mark<\/a>) that give me no reason to trust them and no references? Why, yes!<\/p>\n I want a more authoritative reference. Why? So I can feel more comfortable that other people will follow the same interoperability conventions that I am being asked to follow.<\/p>\n The keyword for the The But wait – when I ask Bloglines to auto-discover the OddThinking RSS feeds<\/a>, it seems to use different titles to the ones I provided in the It seems Bloglines ignores the Unfortunately, WordPress sets that XML field the same for all feeds, so there is no distinguishing marks apart from the URL.<\/p>\n Both WordPress and<\/em> Bloglines could do better here. Bloglines – take advantage of the provided title in the link. WordPress – let the feed’s metadata explain what it contains.<\/p>\n The convention for auto-discovery is to put the links to the feeds in the HTML header.<\/p>\nWordPress and RSS Feeds<\/h2>\n
\n
How do find a feed?<\/h2>\n
Human-Visible Links<\/h3>\n
Orange XML Icon<\/h4>\n
Dealing with Naive Clickers<\/h4>\n
feed:http:\/\/blog address<\/em>\/feed<\/code>. Unfortunately, IE6 doesn’t recognise the feed:<\/code> scheme, and gives a horrible error message:<\/p>\nFeed:<\/code> isn’t a standard URI-scheme<\/a>. Nor does it seem to be used by aggregators doing auto-discovery (WordPress does support the other scheme – described below.) So, my dear WordPress, what was the thinking behind this usability nightmare?<\/p>\nConclusion on Human-Visible Links<\/h4>\n
Auto-Discovery Links<\/h3>\n
link<\/code> tags in the HTML header. The link<\/code> tag specifies the URL of the feed, a title, the MIME type (e.g. application\/rss+xml<\/code>) and a relationship of alternate<\/code>.<\/p>\n<a><\/code> link tags that the web is built upon.<\/p>\nAuthoritative References<\/h4>\n
Where’s Your Grammar?<\/h4>\n
LINK<\/code> relationship is alternate<\/code>. Alternate? Oh please! The word you are looking for is “alternative<\/a>“. Call me a nitpicker, but I find this just plain embarrassing. Did you remember to get someone who has some attention to detail to review your proposal before you published it?<\/p>\nWhere’s Does The Title Come From?<\/h4>\n
LINK<\/code> tag includes a title<\/code> field. That sounds really useful. I have links to multiple feeds in the same heading, and I can clearly distinguish between, say, the articles feed and the comments feed.<\/p>\nlink<\/code> tag. What is going on?<\/p>\ntitle<\/code> field in the link<\/code> tag. Instead, if fetches the RSS feed itself, and looks inside for the title to use.<\/p>\nWhen do you know about your feeds?<\/h4>\n