RSS Feed Classification
August 12, 2005 at 1:06 PM
I have been doing a lot of thinking lately about RSS feeds and how we are reading them and the value they provide. This thinking was provoked once I started trying out NewsGator. I went through the process of exporting some of my web development and design feeds in FeedDemon to OPML so I could import them into NewsGator. It says I currently have 1,177 feeds. So I spent the last few weeks getting used to the new tool.
It turns out that after reviewing most of my feeds, well, a large majority of them are pretty much useless. I have reached the point that subscribing to RSS feeds has become counter productive. The signal to noise ratio is no longer on the side of a strong signal. So I am now pruning back these feeds. But it wasn't until I had them all sitting there in front of me did I realize how much garbage I subscribe to. I realized real quickly that the feeds I really value were getting lost in the waves of general press feeds and feeds for subject areas I am just not into. I was losing control of my content.
It was at this point that I had a thought, what about classifying feeds? What if feeds contained meta data about what kind of content would most commonly be published within that feed overall? Perhaps even meta data describing the kind of content found in each individual post? I am curious what your thoughts are on this, is it good idea or a bad idea?
Just a quick note -- NetNewsWire (Mac only -- http://www.ranchero.com/) supports what they call "Smart Folders," which are really just an extension of the "Smart Folders" supported by Mac OS X.4 (e.g., Tiger). The smart folders search all of your feeds based on certain terms -- subjects, key words within the content, times, authors, etc., etc., etc., -- and it will automatically display feeds that fit your criteria (just like iTunes "Smart Playlists"). Furthermore, since the "Smart Folder" is a permanent thing, it will automatically place new feeds into the appropriate smart folders. In other words, the exact functionality you're looking for, without the hassle of extending the RSS spec... Of course, if you don't have a Mac... I'm not sure what's available.
FeedDemin has the same thing, they are called NewsWatches. This is specifically keyword searching, which isn't really the solution I am looking for. I am told by Tristan Louis there is actually a catagory element at the post/entry level in the spec, but no tools are making use of it, so no publishers are including the elements in their publishing tools.
Oh, and I do have a Mac as well as a Windows system, part of the reason why I was testing out NewsGator.
RSS items can have a category element, which is similar to what you're asking for. FeedDemon enables grouping items by category, but very few feeds actually make use of this element so it's not nearly as useful as it could be.
Sounds like someone needs to start a grassroots organization to advocate the use of catagory ekenents in RSS feeds.
Take a look at the RSS Traits extension. It's a draft that not only can enrich feeds with metadata but also can be leveraged by RSS clients to light up the GUI to help you sift through your feeds. http://www.basement.org/modules/traits
What if feeds contained meta data about what kind of content would most commonly be published within that feed overall? It seems you are referring to an RSS feed, as the result of a search - which, I believe are more powerful than simply subscribing to each site, one by one. The RSS feeds that you really value are a clue to which topics that you enjoy. Each topic is a "personal keyword," of which can be applied to a general search. I've always found that subscribing to feeds, one by one, is one-dimensional. Why limit myself to the topics I enjoy - only on the sites I subscribe to? After all, content is neither black nor white. It's universally informative. I don't care if "Johnny A-Lister" says it, or if "Joe Blog" says it. If it pertains to my interests (keywords), then I'm all for it.
Damn Nick, I totally feel you on this one. I "only" subscribe to 70 feeds right now, and I'm constantly adding/subtracting feeds to make sure the ones I am subscribed to are actually worth the time/effort/productivity loss associated with viewing them. Are you going to be in San Fran for the Blog Business Summit? Hit me up on IM.
1,177 feeds - you must be reading 20 hrs a day! I just checked and I have 41 feeds on my daily reading list. A metadata approach may be helpful - hopefully more helpful than Dublin Core has been for web pages.
Absolutely. One simple piece of punctuation can alleviate this problem ":". The colon. Keyword(s): Topic. There you go. Now all we need is a standardization movement for it. I'm in.
I am not sure what the colon has to offer as far as a solution to this problem. The issue is mainly that the category element, which is a sub-element of an item, is not properly adopted by most blogging tools and feed readers. So an example of a feed might be like: <?xml version="1.0"?> <rss version="2.0"> <channel> [typical RSS 2.0 header goes here] <item> <title>RSS Feed Classification</title> <link>http://www.digital-web.com/news/2005/08/rss_feed_classification/</link> <description>[content of post would go here]</description> <pubDate>Fri, 12 August 2005 13:06:00 PDT</pubDate> <category>RSS Feeds</category> </item> </channel> </rss>
All of my feeds have one or more <category> elements at the <item> level. I did not know so few people used them. Technorati, for example, uses the category element to insert posted items into their tagging system. I also use a category element as an index into the DMOZ taxonomy, although only for completeness, as that is one aspect of RSS syndication I'm pretty sure very few, if any, reader/aggregators take advantage of.
I have always found newsgroups and mail feeds to be horrendous generators of noise. You spend more time deleting stuff you have no intention of reading than actually learning. In contrast, I have found the Live Bookmarks in Firefox to be incredibly useful. This is the only place I use RSS, because the headlines can be easily ignored for a fortnight without needing maintenance. When you are ready to catch up with a particular site, you click on the article that interests you and use its navigation from there.
I used to use live bookmarks, but I found services like NewsGator and FeedDemon much more effective in managing feeds than what options you have in Firefox. Especially when you hit the 1,000+ feed mark.