It’s just data

HTTP Best Practices

Rob Howard: We're going to try and get back to the point where full RSS feeds can be offered. While there are lot of good RSS clients out there, there are also plenty of bad ones that don't honor all the HTTP codes.

To be fair, most of the HTTP headers involved are optional, and most developers are not accustomed to dealing with the protocol at this level.

Despite being optional (which is fair given the range of applications HTTP supports), many of the HTTP features were explicitly designed to solve the problems posed by applications such as syndication.  After all, If implemented properly, the check for new content is an "infinitesimal" burden.

What may help is guides and compliance test suites.

RSS is Eating Me Alive (Almost)

Robert Scoble talks about bandwidth costs of RSS and cites it as a reason more and more sites are not giving full text RSS feeds.... [more]

Trackback from Windley's Enterprise Computing Weblog


So when I first heard about RSS, by first thought was WOW, cool, neat, but shouldn't it also have a event driven distribution model?
Email being the most common model of event distributed data, what if my aggregator ran through my email and pulled out RSS notifications and displayed them in the aggregator.

The power of RSS is XML, HTTP while cool is just the transport.  SMTP on the other had would allow a event based or push model.

I'm not arguing that we should do away with the http pull model, but that there are other solutions.  RSS should have an option that allows consumer aggregators to communicate back an email destination for pushed content.

I currently use bloglines as an aggregator. Bloglines is really a supper aggregator, it pulls feed then your web browser pulls.
Your aggregator becomes the leaf in a hierarchical tree rather that a node in a point to point graph.  Bloglines even gives you an email address to use to subscribe to mailing lists so that messges on the list can appear in your aggregator as news.

SMTP transport could theoretically be replaced with a peer to peer distribution system.

I guess I've described NNTP, well not quite.  But close.


Posted by Kevin Tew at

MS could make things a little easier on themselves by including GZip/deflate support in the .NET Framework HttpWebClient, and making it easy to turn on, or on by default.  My employer has worked hard to get our customers to use Accept-Encoding, but we've ended up having to provide the code for .NET customers to do this, and even then, some haven't been able to get it enabled.  We're doing SOAP, not RSS, but the problem is the same, you can't throw all this stuff on the wire and not expect to take a hit.

Posted by Gordon Weakliem at

RSS Bandwidth Issues

The other day, on Scoble's blog, he announced that MSDN was having problems keeping up with the bandwidth demands of RSS aggregators. Well, if Microsoft can't handle it, then it's definitely a problem, right? Many people have chimed in about...

Excerpt from at

Anne van Kesteren : HTTP Best Practices - Too bad most HTTP headers are optional to support...

Excerpt from HotLinks - Level 1 at

Is there some sort of test suite for checking whether one's own weblog installation follows the patterns - something like the feed validator, but looking not at the content, but at the protocol headers? If not - why not? :-)

Posted by Stefan Tilkov at

Stefan, it is on my todo list.  None of the things mentioned there would be difficult to check for.

Posted by Sam Ruby at

Stefan, the tools at are quite helpful in that.

Posted by Michele Campeotto at

If you must poll, at least poll well...

Microsoft's recent troubles with RSS files have focused the RSS/Atom communities squarely on fundamental problems with feed syndication as practiced today. While some of us have been warning about this problem for quite a while, it has taken Microsoft's recent... [more]

Trackback from As I May Think...


RSS Bandwidth Consumption

Mark Fletcher comments on the recent RSS bandwidth discussions, pointing to recommendations from FeedDemon's Nick Bradbury and also to Sam Ruby's HTTP Best Practices....

Excerpt from Redwood Asylum at

Add your comment