Enterprisey
It started out with a pretty embarrassing post by self-proclaimed industry thought leader James McGovern, and an equally stupid follow-up.
Lots of people followed up on this, including (predictably) Rails creator David Heinemeier Hansson, twice, Microsoft’s Dare Obasanjo, Charles Oliver Nutter, and (indirectly) Jeff Schneider.
McGovern argues that all of those Ruby folks have no clue about enterprise computing, all of the others (except Jeff) claim that if all of those web technologies are good enough for Google, they’re good enough for anything, and Jeff seems to want to keep his options open.
My take is that developing software for real use — as opposed to building it for educational purposes or simply for fun — is hard, no matter which technology you use. Is it possible to build enterprise software using Ruby and plain XML over HTTP? Of course it is, just as it is possible to do so using, say, Java and SOAP over JMS.
“Enterprise”, to me, is a) burdened with the need to integrate lots of existing applications, a decidedly small percentage of which is built using technology that anyone would consider “cool” nowadays and b) much more related to politics than to reasonable arguments. So even if you believe that Ruby and Ruby on Rails is far more productive than any Java web framework — I do —, and even if you believe that building systems RESTfully is The Better Way™ — as I do —, you might end up doing, and even recommending, the exact opposite. As I do.
I have zero doubts, though, that neither Java nor the WS-* stack, as it’s evolving today, will remain largely unchallenged in the Enterprise for much longer.
I think you are right on one aspect of enterprise computing in that it has as a component the need to integrate with other systems in which Ruby doesnt really address. Do you believe that Ruby should step up?
Would love your thoughts on this blog: http://enterprisearchitect.typepad.com/ea/2006/04/enterprise_arch.html