Tim Bray of Sun Microsystems kicked off the day with a talk on all sorts of random thoughts including: standards process and determining success of a standard, the future of RSS, why WS-Eventing is not a substitute for RSS, how atom is here to save the day, and other random comments.


A quick survey of the room tells us that the audience leans more predominately towards the .NET ecosystem.  There is a good representation of folks that work on both Java and .NET platforms and a sprinkle of people that deal with Java only.


The most interesting link he pointed to during his talk is on the state of blogging and RSS as the #1 application of XML: http://www.sifry.com/alerts/


Tim's speaking style can be hard to follow.  You have to often infer where he is going with his talk making assertions without backing them up leaving you to fill in the holes.  I did like his approach to his talk by using a web page to point to a series of links where he spoke to content on each.  This helped after the talk to bring context to some of his assertions.


Coming from a Unix heritage, Tim thinks about data first and not APIs.  Why map your data to an OM?  Just send messages around and parse XML cause it is simple.


I wonder if people working with HL7 Documents feel that way?


Tim feels we need to learn from the past when designing standards.  When he was designing XML, during the end of the design cycle, they pushed back on people who tried to randomize them with complexity and new functionality stating it would be in v2 of the standard; however, v2 never abounded and will likely never come to life.  Design your standard with the assumption that there will be no v2 and keep the standard confined to the smallest subset of functionality for only the required and needed scenarios today. "If you don't have an urgent burning requirement for things right now.  You ain't going to need it.  Failed standards have too much stuff in them."


Everyone generally agrees that Schema has it's shortcomings, but it is facilitating huge amounts of value for customers.  Thousands of business documents today are based on a schema.  So, is XML Schema a failure?


Can simple standards scale to solve bigger problems?


Why ATOM? Is it not simply v2 of an existing ubiquitous standard, RSS?  (I am assuming that RSS v2.0 as it exists today is really v1 in reality since it is what was adopted)  It will be interesting to see if the blogging world's tooling is so well ingrained that it becomes a challenge to innovate in the blog space especially on the level of trying to design an entirely new architecture.  I believe that RSS has already exploded and exists.  I am not sure if it will be possible to shift this boat.


Tim talks about a customer back in the day looking at SQL and stated, I can't possibly put my complex data into a square table.


Did the customer know what they needed to solve their problems?  Sounds like there wasn't a burning need for the technology in their minds, but as it turns out the innovation ended being overwhelming successful.  When do people really understand the impact of a standard?


Lot's of great discussion here!


I have a reminder to Ask Tim tomorrow if he was happy to be here - I wonder if he will let me film him while he answers that question.