Day one’s bang continued through the energy level picked it up a notch as keynotes and general sessions began. Live Clipboard announcement made by Ray Ozzie was quite interesting. The Multi-touch interface presentation was just plain cool. Lots of coverage over on here.
On a more mundane, but still personally enlightening moment during the action I had a chance to hear Sun’s Tim Bray talk about standards development. Having lead successful efforts to standardized XML and more recently Atom, he has a unique and accomplished resume that few in the Internet world can compare.
He said the most important thing for a standards body is to not invent technology. Ideally, standards should be based on a substantial body of experience and prior art. The job of a standards working group would be to write best practices that work.
Why try and replace something so successful RSS? There are problems with the format as used today. Further complicating matters was the contentious politics surrounding the format and the fact that RSS 2.0, the most popular form for the format, had been declared frozen despite these problems. He noted that “the syndication community has always been a loud and fractious noisy place.”
The desire to evolve syndication and address the problems in a contentious community necessitated a new and separately named effort to be created. This eventually this lead to an IETF work group to be formed. Whether that was the right thing time will tell Bray said.
“The IETF brought a formal process that no one can claim that they were left out or ignored.” The noted that the core group of Atom members could have sat around in a basement for a week or two and gotten pretty close. It wouldn’t have been as good though.
Approximately 2 years after the effort began, the Atom Syndication Format has been named an official IETF specification. Currently the Atom Application Protocol (APP) is in interoperability testing and is getting close to a final call he reported.
The working group comprised members with a great deal of experience with syndication. What the IETF also brought was a formal process and great deal of know-how security. He later said “We’re probably going to get the crap kicked out of us when we send the APP around [for IETF review]. I think that is a good thing.”
In order to lead such efforts you have to have a thick skin he quips with a smile. He said the toughest part of being involved in these efforts is the “violent personal attacks” that will crop up. He notes that there is something about syndication that “flips the a**hole bit” in normally nice and intelligent people.”
Later in the day during his session presenting Atom as a case study he lampooned the two years of discussion and debate with a slide show of various depictions of war, carnage, and destruction with the Flight of the Valkyries playing.
Devising the technology is not the hard part, dealing with people and forming consensus amongst them are.
“RSS 2.0 is fine and its going to be with us for a very long long time” he said, noting that despite its problems it is good enough for blogs & straightforward news feeds. RSS2 causes real problems for more “technical” feeds.
In his presentation he listed drilled down on the problems you can run into:
- Enclosures - One? Or More?
- Silent Data Loss - Depending on an aggregators assumptions sections of content will silently disppear.
- Punctuation Pain
- Relative Links - no affordances are made making publishing more difficult and documents more verbose.
- International Resource Identifiers support - all links must be ASCII.
- APIs — The MetaWeblog and Blogger API map poorly to RSS and are highly under specified and BAD (broken as designed).
- Aggregators misinterpreting the display punctuation characters
Bray believes the Atom protocol will have a bigger impact then the syndication format. He said, “the APP is the missing infrastructure link in making the Web writable by everyone.”
The group will probably turn its attention to the auto-discovery specification. He expects the working group to disband by the end of the year.
He said he believes it will by highly beneficial to the world if there is no Atom 1.1. The Working Group was gone to great lengths to devise a well defined and powerful extension model for anyone to build on. What he expects to see is more specialized extensions be developed and standardized in varies forms. Efforts are already under way.
Bray’s posted his case study links here.
