A topical political row in the UK between church and state reveals some underlying issues for the service economy.
The UK Government has decided that the adoption process should not discriminate against gay couples, and relevant legislation will come into force in April. But a portion of the adoption process is outsourced to faith-based agencies that are not willing to provide such services to gay couples, and these agencies have requested special exemption from the legislation. (Protests have been led by Catholic adoption agencies, with the support of the Roman Catholic hierarchy in the UK, but other Christian leaders have also expressed their support.) [BBC News]
I do not wish to comment on the issue itself. I want to discuss some more general structural questions.
Firstly, we might look for a way that the Catholic agencies could continue to provide some valuable services in this area, that is compatible both with their religious beliefs and the prevailing legislation. It might be possible to decouple the contentious parts of the adoption process (selection?) from the non-contentious elements (support?), and find ways for the Catholic agencies to provide the non-contentious elements in a mutually acceptable way.
This would presumably involve changes either to the way the adoption process is decomposed into services, and the way these services are funded, or to the ways in which different agencies collaborate with one another to deliver a complete set of services.
This raises a second important question. If a service provider has a strong religious or ethical objection to X (whatever X may be) then we obviously don't expect that service provider to provide services that perform X. But what about delivering services into a larger composite system that happens to include X?
Thus the Catholic adoption agencies may be faced with a difficult choice. Either to provide some useful services, even though these services may be used within a larger system that includes elements the Catholic agencies don't like. Or to refuse to provide these services unless they are able to control the end-to-end process according to their own values - perhaps in the hope that their protest will result in some larger change in public policy.
Note that we don't always know how our services may be mashed-up or reused (repurposed). P.G. Wodehouse once wrote a story in which Bertie Wooster and his friends ran a gambling syndicate on the length of sermons in country churches. Presumably the clergymen were not aware of this syndicate, and would have been horrified to find out. But they probably wouldn't have stopped delivering sermons just because some people were abusing them.
On the internet, of course, you may be able to monitor how people are using your services, and close down or inhibit inappropriate uses. But how feasible is this, given that it is sometimes difficult to take any action without causing problems for legitimate users?
The more general point is that an open network of services will almost always involve heterogeneous value systems - whether this is a secular public sector working with faith-based voluntary agencies, or IT vendors working with an increasingly complex ecosystem. Just yesterday for example, Chandler Howell discussed Apple's growing difficulties imposing secrecy on its partners [Secret Apple Sauce].
These are (among other things) structural problems. The SOA architect may not be in a position to resolve value conflicts, but may be able (in collaboration with others) to produce structures that can operate effectively and efficiently despite the (inevitable) value conflicts. Open or closed.
Technorati Tags: SOA service-oriented