Bill Burnham's posting (no permalink), 'RSS: A Big Success In Danger of Failure', with its conclusion that 'without the use of meta-directories and their standardized classifications and taxonomies the RSS community is in danger of collapsing under the weight of its own success', has attracted attention. It appeared in The Weekly Read (permalink), an on-line publication of Seed Capital Partners, an early-stage venture fund. The comments at The Weekly Read contribute to the debate, but of particular note elsewhere are Scobleizer (where Dave Winer has a comment) and Don Park (who agrees with Burnham). There are good postings, too, from David Galbraith, Jon Udell and Robin Good.
Other notable postings concerned with how to use RSS efficiently and intelligently include this from Nick Bradbury:
Before long, reading feeds starts to feel just like reading email - you look at each unread item just in case it's important, but most of the time there are only a few items that really interest you. Eventually it all feels like work, and you stop enjoying it. But this isn't email, folks - you don't have to care about every last item you receive in your feedreader, as long as your feedreader provides enough help. The most obvious way that FeedDemon helps is through its watch feature. A FeedDemon watch examines every incoming news item, but collects only the ones that contain specific keywords. This way you can subscribe to hundreds of feeds, yet only be alerted when one of those feeds contains something that interests you. ... FeedDemon's newspapers help even further by grouping news items. Although many people choose to read their feeds one at a time, others may prefer to use FeedDemon's group newspaper so that they can scan all items from a group of feeds at the same time. ... As more people become avid feed readers, the problem of managing all this new information will become even more important. I designed FeedDemon 1.0 with this in mind, and this will be a primary focus (if not the primary focus) of future versions.
Finally, there is this from Bob Wyman (of PubSub):
The solution to the problem comes through focusing on an often overlooked element which exists in many computer-based categorization or taxonomy systems. In Topic Maps, this element is the Subject-Indicator which is a URI that identifies a "Subject". A Subject is "anything that can be spoken about or conceived of by a human being." Topics, which are organized into Topic Maps or taxonomies are simply slots or labels that define the position of subjects within the taxonomy. Thus, one might have a subject such as Shakespeare's Hamlet, which is identified by the unique Subject-Indicator "http://example.com/Shakespeare/Hamlet" and a Topic "Plays.English" which is defined to encompass discussions that refer to the subject-indicator "http://example.com/Shakespeare/Hamlet". The important thing to notice about a subject-indicator is that it, like "subjects" in the real world, is not defined in terms of a taxonomy. Rather, the taxonomy is defined in terms of the subjects it covers. Of course, this maps to reality. We can have subjects of discussion even if we have no idea how to classify them -- which is pretty much the situation we're always in on the Internet... Even if we know how to identify a subject in one taxonomy, we often have no idea how to identify that subject in some other taxonomy. In fact, we usually have no idea what taxonomy will be used by our readers. Will they agree that Hamlet belongs in the "Plays.English" topic, or will they insist that it belongs in "authors.english.shakespeare.plays"? While there will be much debate about the topic, there will be much less debate about the subject. Hamlet is Hamlet. And, "A rose by any other name (or topic) would smell as sweet." If we can identify the subject or subjects of a posting in a taxonomy independent fashion, then any number of taxonomies can be defined which define the place of specific subjects within those taxonomies. It then becomes possible to ask for all items that discuss "Plays.English" or "authors.english..." and receive links to Hamlet without needing to translate between or agree on common taxonomies. The taxonomy which is used in a search becomes the choice of the searcher -- not something that the author of the searched-for text must agree to and reference in their text. Of course, some readers will recognize that this basic insight is not a new one. For instance, the "Easy News Topics" proposal that came from the ThreadsML discussion incorporated this insight as have other proposed systems. PubSub.com offers specific support for working with subject-indicators in its ability to subscribe to messages based on their "Referenced URI's". This facility, available on the advanced subscription pages for Weblogs and NewsGroups, allows a user to subscribe to only those new items that reference, either in the text or in hidden tags, any URI or URL that might be interesting to the user. Thus, it becomes possible for authors to insert "subject-indicators" into their postings and have others find discussions of those subjects simply by subscribing to the subject-indicator URIs. Others who wish to write on the same subjects can simply refer to the same subject-indicators in their postings and their messages will be easily found. One can then build taxonomies that define the classification of various subject-indicators. The result is the ability to search by topics without requiring the author of a posting to even know that the taxonomy exists. ... By focusing on the much neglected problem of identifying subjects in a taxonomy or classification independent fashion we can make it much easier to address the problem of classification. The key is to identify subject *in* the published item while dealing with classification *outside* of the item. Subjects should be globally identified while classifications should be determined locally, personally, or in a task-appropriate manner. Only in this way can we hope to be able to address the problem of classification without forcing everyone to accept the same classification scheme.


