On Wed, Mar 26, 2008 at 8:07 PM, Steve Jones <[EMAIL PROTECTED]> wrote: > I'm not so sure about that. Business Services may not suite REST that well
Hang on a minute, are you saying that BS may not suit REST simply because we've got lots of WS-* already? Or is there something about REST that makes the more "enterprise" operations difficult? Actually difficult, or just because big vendors ain't doing it ... yet? > The point for me is that there are different types of services some > (especially those in the interaction space) are probably best served by a > REST type approach. Those in the optimisation or transactional spaces are > probably best served by a WS-* approach. But few will choose to intermingle both. People have a tendency to think there is one holy grail, and use that hammer on anything that looks like a nail, right? I suspect, though, that we'll soon see middleware that does REST / WS-* translations, which might be an interesting area to watch. > The important bit for me is that > what we need is to clearly understand how RIA (a presentation approach) will > interact with users and how that interaction will be finally represented as > transactions in the business. This isn't a single unified stack, its about > how the network of services in the business can be leveraged. Hmm, as REST in a SOA - often referred to as ROA - promotes a web of resources more than service start- and endpoints, I'd assert that REST is *more* like a network that most ESBs (or any other similar stack) are these days. > For me there is no RIA or Mashups at a large scale unless there is effective > SOA in the enterprise. It doesn't overly matter how those services are > implemented as long as they represent strong business services which can be > combined in the way that the business needs. But what is a strong business need? Very often those words are used when you want to expand markets, test out ideas, try out something new. I'm sure there's always the small prototype to test out some small idea, but I'm often involved in mashups and prototypes that are *huge*, testing out not only frontend stuff, but middle and backend as well, all the way up to the very business idea of the organisation. I can understand the need for a solid infrastructure (of which it sounds to me like your SOA belongs) to support the *current* business, but it is vitally important to not make it too solid. I thought that was the whole point of SOA; agility across business thinking *and* implementation, no? Alexander -- --------------------------------------------------------------------------- Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps ------------------------------------------ http://shelter.nu/blog/ --------
