Fatten or strip - the great Java debate

Meijer called for prudence when architecting a language at the very beginning, as you can hurt legacy infrastructures and break interdependencies by removing APIs later on. "We have to be more careful about puling things out. Once it's in, you never take it out," he warned.

On the other hand, Bloch supported adding "good syntactical sugar" that doesn't Increase the "surface area" of Java for students and academics.

For Bloch, that meant saying "no" to Java Specification Requests (JSRs) for things like XML literals but - surprise - "yes" to things like a forthcoming Google-sponsored JSR Bloch said would catch multiple exception types to help simplify programming. That left some in the audience who wanted nothing more than some sweet XML-like syntactical sugar feeling distinctly sour.

Bloch found backing from panelist and Sun JRuby developer Charles Nutter, who said Java should "solve pain points of developers in a measured way, not in a C# way where you throw everything in or Perl 6 language where it's the 'every language'." His colleague, Chet Haase, an architect at Sun's Java Client Group, suggested Java could be made easier and appealing to a larger segment of the population - such as consumers - through changes to the language or libraries.

So just whose to decide what's in and what's out? The community. Ah, the wisdom of crowds.

"We should ask the users [what they want] rather than making assumptions," Johnson said. "[We] shouldn't assume people don't want to make some effort if they see it's worth it in the end."

And that's where Sun's OpenJDK could come in, according to Nutter "There's noting to stop anyone developing their own version of the OpenJDK. As we see more and more of that, it's going to feed back into the [standards] process, and we will know if people want us to pull things out."®