Well the shiny kids have a new toy... its WebSockets and by jiminy if it isn't just the best thing ever. Actually to be fair to people promoting WebSockets they do appear to be taking a more rational, and less religious, stance than the REST area but this discussion remains pointless. Mark Little's recent InfoQ post was a good pitch on the debate.
Just as REST v WS-* is pointless so the WebSockets debate is pretty pointless as well. Its a new way of shifting information from A to B, its an 'improvement' over REST and WS-* in some ways and not in others but it doesn't actually make the job of conceptualising solutions any easier.
Its a rubbish picture but it gets the point over. REST, WS-*, WebSockets, IIOP, MQ, Flying Pigeons, Rats with backpacks and any other data shifting technology are about providing the mechanism via which two services can communicate. Its a required piece but the key question is 'what is the cheapest way', the value is provided by what goes across that mechanism and to define that you need to understand how the producers and consumers need to interact and that requires a conceptual model and thinking.
The hardest part of IT remains in that conceptual part, the planning, the thinking and the representing back to the business what is being done. REST, WS-*, WebSockets or any other mechanism do precisely bugger all in helping to get that done. The question I'd pose is this
Its 2012 now and what has improved in the last ten years in terms of making systems easier to conceptualise and the business concepts easier to communicate and what has been done to make the translation of that into technology (producer, interaction, consumer) much simpler and straight forward?
From where I'm standing the answer appears to be a whole heap of bugger all. Does WebSockets make this problem much easier or is it another low-level coding mechanism? Its of course the latter. When people moan about the walled garden of Apple or 'monolithic' ERP they are missing a key point:
Just as REST v WS-* is pointless so the WebSockets debate is pretty pointless as well. Its a new way of shifting information from A to B, its an 'improvement' over REST and WS-* in some ways and not in others but it doesn't actually make the job of conceptualising solutions any easier.
The hardest part of IT remains in that conceptual part, the planning, the thinking and the representing back to the business what is being done. REST, WS-*, WebSockets or any other mechanism do precisely bugger all in helping to get that done. The question I'd pose is this
Its 2012 now and what has improved in the last ten years in terms of making systems easier to conceptualise and the business concepts easier to communicate and what has been done to make the translation of that into technology (producer, interaction, consumer) much simpler and straight forward?
From where I'm standing the answer appears to be a whole heap of bugger all. Does WebSockets make this problem much easier or is it another low-level coding mechanism? Its of course the latter. When people moan about the walled garden of Apple or 'monolithic' ERP they are missing a key point:
Technical excellence in the mechanism isn't important, its excellence in delivering the value that counts.See you in another 7 years for the next pointless debate.
1 comment:
A Whole New Way to Think About Service Oriented Development covers how to consume and use SOA from websockets (or starts to cover, more coming). Websocket based SOA has some unique characteristics and some unique challenges. Take a look at this, and then contact me. I want to pick your brain after a few more posts and a slide or so more.
Post a Comment