As a technical guy, I don't get it either.

• Posted by: Alex Moskalyuk
• February 12, 2004 10:22 AM

The key issues are:
1. There is no "one" RSS. In fact, depending on how you count it, there are between seven and nine different versions of RSS.
2. Each of the many versions of RSS is, in some way, incompatible with the others. Sometimes the difference is very large. For instance, some versions of RSS look like "normal XML" while others are in an RDF encoding. Thus, building an "RSS" processor means building a tool that processes not one format but many.
3. The RSS specifications (when written versions actually exist...) rank among the worst specifications written for any Internet related tool or format. They are grossly underspecified and filled with ambiguities or important use cases not addressed.
4. The most prominent evangelist for RSS, Dave Winer, has declared that "RSS is finished" and there will be no effort made to address any of the existing failings in the specifications nor issues that may arise in the future. (Note: I've been developing distributed apps for over 25 years now. RSS is the *only* spec that I've *ever* heard anyone say would never be modified -- unless it was for a dead product. By declaring that the spec is closed, Winer has declared that RSS is dead...)
5. RSS, in Winer's versions, ignores XML1.0 standards in a number of ways. For instance, these versions include a tag that is to be used instead of the xml:lang tags which are standard in XML1.0. However, the definition of the RSS language tag results in massive difficulties in handling a number of I18N problems, like feeds that contain items written in more than one language.
6. Date handling in RSS is poorly defined. There is only one date (pubDate) when at least two are required (created/modified, issued). Also, dates are in the old RFC822 format (i.e. "Thu, 12 Feb 2004...") rather than an international standard format.
This list of failings in RSS could be extended at great length. The point here is that RSS was a wonderful idea created by Netscape many years ago. But the RSS specification has been badly maintained over the years and now desperately needs updating, a proper specification, and an open process for resolving issues. Atom is intended to provide RSS style syndication with the solid, well defined base for interoperability that it deserves and requires.

bob wyman

• Posted by: Bob Wyman
• February 12, 2004 11:00 AM

Be flip, be very flip -- just as flip as we all were about the difference between HTML 2 and 3. From the commercial point of view, the RSS/ATOM argument is just a market-growth impediment that is best ignored in public.

The technical guys are all correct in arguing the points out regarding the metadata inconsistencies/omissions in RSS and the security hazards in ATOM, but Feedster's and everyone else's coders are just writing parsers for all the different formats already. It's annoying, but no more than that.

When the aggregator market consolidates and the black hats start exploiting ATOM holes, the variance among the data standards will naturally decrease as fewer, more robust products get all the market share.

• Posted by: Scott Rafer
• February 12, 2004 11:28 AM

Pirillo says it even better: http://rss.lockergnome.com/archives/opinion/008978.phtml

• Posted by: Scott Rafer
• February 15, 2004 8:14 AM

if you want to *publish* eg. stuff from a mailinglist to people who just want
"the news", the question is what format to use! everybody else just sits and
waits or uses (a variant of) mark pilgrims universal parser (python).

• Posted by: clemens fischer
• November 18, 2004 8:50 AM

if you want to *publish* eg. stuff from a mailinglist to people who just want
"the news", the question is what format to use! everybody else just sits and
waits or uses (a variant of) mark pilgrims universal parser (python).

• Posted by: clemens fischer
• November 18, 2004 8:51 AM