The Future of RSS

Highlighted by edutechie

Highlighted by edutechie

Highlighted by omonad

Really Simple Syndication

aggregators

Highlighted by omonad

RSS in a Nutshell

RSS is rather simple language to describe the latest headlines (or the full content of articles).

Highlighted by omonad

Highlighted by omonad

Highlighted by omonad

RSS - Beyond the Distribution Medium

So today RSS is a great distribution medium. Why? Because it has become ubiquitous. If you are an online business with customers and you do not utilize RSS, then you are simply missing out. Smart companies are leveraging blogs, photos, video, podcasts to stay in touch with customers daily. Other services, like del.icio.us (owned by Yahoo), allow users to publish and subscribe to feeds, enabling powerful social networks outside the website.

The ubiquity of RSS is so powerful that publishers want to deliver more and more content to users via RSS. But the problem is that basic RSS cannot be used to deliver structured information.

Highlighted by omonad

Highlighted by omonad

We need something lighter and more portable to carry our information around - hence XML and RSS. 

Note that businesses are probably the most interested party here, because to a business a loss of structure leads to loss of meaning, loss of trail and ultimately the loss of customers.

Highlighted by omonad

Extending RSS

To extend RSS basically means to add a custom tag. For example, Google Base currently has 148 attributes that it recommends to add to RSS. Here are some examples starting with the letter 'a': age, actor, agent, apparel type, artist. These are everyday concepts that might come handy in classifieds and other aspects of life. All of these tags allow Google Base to make RSS structured, whilst preserving its basic capability. 

Similarly, FeedBurner inserts proprietary attributes into their RSS feeds. This is done purely for house keeping purposes, because only FeedBurner's engine is meant to process these attributes.

The main problem with extending RSS is agreeing on what things mean. In the case of FeedBurner it is not critical, but in the case of Google Base it is much more important. In order for RSS extensions to work, the second piece of the old technology dilemma needs to be solved. There needs to be a common format for communicating data between applications:

Highlighted by omonad

Conclusion

Purists, myself included, would argue that using RSS for the delivery of complex content is a hack. After all, what does a news format have to do with semantics? But technologies do not evolve in a pure way. Some things catch on and succeed, and become widely adopted. The fact is that RSS is becoming a pervasive on-demand technology, which outweighs the fact that it was never meant to be the semantic agent of the web. But even from a purist's perspective, there may not be much to pick at - RSS is just another XML-language and in that respect it is as good as any other flavor of XML.

So will RSS become more than it is today? Will it be able to solve the second piece of the old technology industry puzzle - the common format? As usual, only time will tell.

Highlighted by omonad

Please... Do not extend RSS any more!

For god's sake (place any god here), RSS should NOT be extended anymore. The list of current formats (that in the end do the same thing) is HUGE and already really complicated with incompatible formats with one another:

RSS 0.90, 0.91, 0.92, 0.93, 0.94, 1.0, 1.1, 2.0, 2.0.1
ATOM 1.0, 2.0

See a "very" old post on this subject [The myth of RSS compatibility] http://diveintomark.org/archives/2004/02/04/incompatible-rss

RSS means many things... but lately the most common meaning for it is: really SIMPLE syndication. SIMPLE as in: NOT totally complicated as it already is now.

RSS is meant for people that use aggregators to get the content of internet sites. It is not meant to be used as mean for transferring data. This is also already solved! It is called XML and is used to carry data, even execute bank transactions... whatever.

RSS is a subset of XML... If you are asking for RSS expansion it is like you are asking for RSS to become XML... Why not just invent a new name for it like "TDML" (transfer data mark-up language) and use this. RSS was never meant for carrying data for bank transactions or anything like that.

Highlighted by omonad

This really is silly. For domain-specific information, we've got XML and RDF. Why use a content syndication format to deliver domain-specific information when we have technologies like XML and RDF which are far better suited for that purpose? RSS can be extended for simple things, but there becomes a point where it becomes easier just to use the technologies specifically designed for the purpose of data interchange.

For instance, take the problem of data persistence. If you are using RSS, the data stays in place in the RSS feed only as long as it's fresh. Typical RSS behaviour is after fifteen entries, the information disappears from the feed. Podcast archives are a problem for this reason - if you want to go and get an old episode of a podcast, there's no structured information. With generic XML and with RDF, data persistence becomes a lot easier to deal with - because one can, for instance, use a URL to point to previous entries.

Highlighted by omonad

Highlighted by omonad

Highlighted by jackie

Extending RSS might seem like a practical way to address these very real problems, but as you say, RSS took off because it was Really Simple. The more complicated you make it, the further it strays from what made it attractive in the first place.

A big problem with all the 'structured' data integration approaches is that no one can ever agree on what the structure ought to be. Look back to all failed attempts to solve this problem, from EDI, cXML, x12, etc. Yuk.

Highlighted by omonad

The banking/invoice example makes me think that it isnt RSS that needs to be more structured, it is instead a need for smarter and more flexible XML readers. Note that I say XML, not RSS specifically. This is because what I feel is needed is focus on having a system to register custom XML DTDs that xml readers (web or client) can then understand (machine-readable compat). When I say register, there should be a way for X Bank to become associated/binded with a given XML DTD. The Reader software just needs to parse out the XML structure and present the data... Include rules and parameters to the system and things get even cleaner for the benefit of the user viewing the data.

So, it's not so much about RSS adoption as it is just making use of XML for a plethora of business needs. RSS serves well for a rather broad market, from headline news to video/audio/image distribution. But this doesnt mean that RSS is the right spec for everything.

Highlighted by omonad

For simple content delivery, RSS 2.0 is ok-ish though it is broken in a few places (such as markup in titles), and it's not very well defined for extension. Atom can be seen as a bugfix version, and can be pushed a lot further.

As Tom suggests there are better defined approaches of representing and delivering arbitrary data, in particular RDF. But I don't think it's really an either/or situation. Atom is nicely defined for content syndication, and has some nice features in place such as entry versioning which could be used to benefit in delivery of broader data.

The Atom Publishing Protocol is going to open a lot of doors, very much allowing Read/Write Web using the standards as originally designed. The extension potential there is also pretty good. It remains to be seen where this will lead, but it's likely to be very good for the web. As a Semantic Web fan I like that - what's good for the web is good for the Semantic Web, and vice versa.

Highlighted by omonad

"The main problem with extending RSS is agreeing on what things mean"

RDF has means of doing this through the creation of simple ontologies which declare equivalence. So, if I created an ontology of all English words (just to give a purely academic example that would never work in reality), I could specify that when I say "trousers" in my English namespace and someone writing in the American namespace uses the word "pants", we both mean the same thing.

The RDF/SemWeb people have been working on this for a long time. It is a tremendously simple way of merging large sets of data from disparate sources. But, as usual, everyone will write it off as being too complex and have all the usual naming conflicts even though URIs and namespaces can solve the problem...

Using RSS for anything but the most basic of data interchange seems to be a case of reinventing the wheel.

Highlighted by omonad

Highlighted by omonad

Highlighted by omonad

Highlighted by omonad

Alex,I think matthew is right. RSS is not a push technology. It seems that the new content is push to the consumer. But actually, it is the client side of the technology doning the job to pull the new content from the web site serving the rss feed.

The difference between push and pull is just like the postman send you a newspaper everyday and you buy it youself.In fact,subscribing a RSS feed by giving your email to the website, that is a push technology, since the host of the website konw your email address and push the new content to you when they are avaiable, just like the postman know your address.

Highlighted by omonad

Highlighted by joethink

Highlighted by drthomasho

Highlighted by drthomasho

Highlighted by jackie

Highlighted by jackie

Highlighted by jackie

Highlighted by jackie

Highlighted by jackie

Alex is right, RSS is conclusively solving the delivery problem. The high-level concept of a sequence of dated items is applicable to many problems and that is why RSS is attractive - there's no reason to invent anything different. RSS is a great wrapper for all sorts of dated information.

The format question is really about what lives inside the description element; there are two sides to this. Not all feeds have to be consumed by a conventional feed reader, the date wrapper that RSS provides is useful for machine-to-machine applications. On the other hand, XML (including RSS extensions) was always intended to be human-readable with the tags ignored so properly structured data should still render in some way.

So for a banking application for example, RSS would provide a convenient transport to get transactions or statments into your software and as a by-product lower the bar for other developers to do useful things with it. This is of benefit to users without them trying to consume the feed in the usual way, although if they did it should still make some kind of sense!

Highlighted by omonad

Highlighted by melissa

Highlighted by drthomasho

Highlighted by drthomasho

148 attributes

Highlighted by wisely

Tom Morris

Highlighted by wisely

# 21

Good article, but I would replace "RSS" with "Syndication" all through it. RSS is deprecated. I can no longer see why we're still hanging on to it. Atom is currently the best (if not only) alternative.

Posted by: Henrik Lied | April 4, 2007 10:00 AM

Highlighted by wisely