I have recently discussed my skepticism over the use of mashups as , but there isn't much question that such products are well in the offing and will be pushed hard on businesses in the near future. With the increased visibility and marketing hype, Joe McKendrick reports that there is some pushback from some Service-Oriented Architecture proponents in the matter.
Briefly, the thinking is that mashups represent quick and dirty methods with which to achieve some of the typical benefits of an SOA, without putting in the spadework and laying the foundation properly. As with all things done in such a manner, this would have implications as to the long-term efficacy of the approach as systems grow and encroach. McKendrick points out, however, that the visible, immediate utility of mashup tools (if indeed this exists, as alleged) may be the best selling point possible for introducing non-technical executives to the advantages accrued by adopting SOA.
For my part, I don't see the tension. While it's true that some mashup tools are being sold as basically SOA in a box, when delivered appropriately, mashups should simply be a component making use of a properly built SOA, not some sort of shortcut over the architecture. I have no doubt that some people will try to use them in place of building out a reliable service-oriented system, but people are trying to do all sorts of things to achieve SOA-like results without actually investing in SOA, so you can hardly blame mashups for the issue.