What I usually do in this space is grumble and whine about SOA & SOAP and so on. Today, let’s start with a laugh, instead. But actually, there’s news; the Java tribe has decided to take REST seriously, see JSR 311 and, for some more pointers, Eduardo’s write-up. I haven’t had time to do a deep-dive, but I’m reassured by the presence of Marc Hadley A.K.A. the WADL guy; and it looks that there will be other real experts at the table. On the other hand, Elliotte Rusty Harold emits an extended snarl. (Sample: “I hope we can derail this completely...”) Hey Elliotte, I guess making friends and influencing people is for losers, right? The proof of the pudding, obviously, is in the eating, but the fact that this discussion is happening has to be a good thing.
Update: the RESTafarians are piling onto the expert group, and Marc Hadley shows some sample code.
Comment feed for ongoing:
From: Daniel Haran (Feb 14 2007, at 14:34)
Any chance the Java RI will be close to what the Rails crowd did with Resources?
http://rails.rubyonrails.org/classes/ActionController/Resources.html
[link]
From: Jason R Briggs (Feb 14 2007, at 16:09)
Have to admit, I agree with Rusty on this one. Hard to see this turning out well. I'm prepared to be proved wrong, but I'll remain safely pessimistic about another design-by-committee effort producing anything other than complexity.
[link]
From: Derek Kaz (Feb 14 2007, at 19:36)
Sample code looks hideous...gotta agree with Elliott Rusty on this one
[link]
From: Simon Brocklehurst (Feb 15 2007, at 00:31)
I guess some of the negative reaction is understandable: after all, historically, major players have had a vested interest in building unnecessary complexity in to standards. They could get away with this, because they knew they would be the only game in town.
However, the software world looks very different in 2007, compared to the way it did in the early 1990s - the big different being a really strong Open Source movement. One of the great things about the Open Source movement being so strong is that now, these same big players have to compete in an environment where high-quality Open Source options exist. Today, when "official" standards aren't up to scratch, developers simply use other frameworks.
I hope this new spec turns out to be great - there could be genuinely useful benefits to having a Java REST standard, not least making great tool support possible (e.g. in Netbeans).
[link]
From: Henry Story (Feb 15 2007, at 01:26)
I am happy to see Jerome Louvel as an expert on the group. He is the author of the RESTlet Java API, which is real running code. But I don't yet understand what JSR 311 is promising. A good RESTful framework in Java would certainly be great. But why not start by standardising RESTlets?
[link]