Legacy costs come in many shapes and sizes, and the more I look at them, the more I realise how much of an albatross they can be. They really obscure people’s vision and can corrupt decision-making unless care is taken.
- It is said that the first $3000 spent on building a US car goes to retiree healthcare, making the industry less competitive.
- During the run-up to Y2K, I was asked to look at the preparedness of some Polish banks; their systems, post 1990, were too new to have the problem. It’s the banks with older systems that had to spend money checking everything.
- Many firms (like the one I work in) capitalise software development, creating a sins-of-the-father problem for later years.
- Having lived in Calcutta and in London, I have seen first-hand some of the problems of being “the first city” to do something, particularly in transportation or utility. You can’t get rid of it as easily later.
- The spaghetti of regulation covering two areas, telecoms and intellectual property, behave similar to tax regulations. They are complex and people exploit the complexity and there are winners (who would do anything to hold the status quo) and losers (who will do anything to change it).
- The minds and hearts of the people used to the legacy.
Regulation and bad law. Infrastructural investment requirements. Accounting policies. The difficulties of having to perform complex change operations on “living” things, much like skin-graft meets open heart surgery. Human inertia. So many legacy costs.
It is so much easier to start from scratch.
That’s what I used to think.
Until I saw more and more of what today’s emerging technologies could do.
Step 1, in the mid 90s, was to wire up static pages and queries. Step 2, in the late 90s, was to wire up transaction initiating pages and forms. Step 3 was to connect these things together and form mini business models.
But step 4 is to allow co-creation. Not just co-creation between people, but between applications. SOA allows for co-creation if we understand the power of things like search and syndication. We need to rethink data mining and EAI and all that jazz, we need to understand that the marginal costs of storage are trivial, that we can afford to record everything and tag everything.
Twenty years ago, if I said that the most important component of a trading system would be a Visicalc lookalike, people would have laughed at me. Now even though I want to, I cannot get rid of the stuff.
We need to understand the value of syndication without being bogged down in all our yesterdays. Legacy does not mean albatross. But we have many Ancient Mariners.