This is a slightly updated picture from one I posted on twitter the other day. It’s an attempt to express my general view on high level “should be possible” integration conversations that get misunderstood as “clearly simple and straightforward and definitely happening”.
In my mind I’m arguing against tightly-coupled integrations between disparate systems that do not have OOB or proven integration, so we’re custom building inter-dependencies that will require a lot of deep knowledge, of both systems, to maintain over time. However as @JayVanDerZant pointed out on twitter, even the genuinely good ideas can get side-tracked in the excitement of this “blue sky thinking”.
So lets all be practical. Yes we probably can do whatever’s being discussed, it’s undeniably possible, but once we think about budgets, risk and the long term costs, is it genuinely a good idea?