Went to a mandatory AGILE methodology meeting session currently being pushed by the client. Personally I consider myself already sold on Agile methodologies, so feels somewhat that it was going to be a gigantic waste of time. But it was mandatory, so to keep everyone happy, I went.
I went with a notebook, and just moments before the session, I had a brilliant idea of noting down any new words I'm about to learn that day:
We call it a deep-slice, but it's slightly different. A spike solution is a focused technical effort to explore whether a technical design is actually possible or feasible. In the particular Agile session it seems the spike solution is also a throw-away solution. But I can see it being documented on a dev-wiki until it's used in implementation.
A spike-solution is good because it reduces risk - if an approach can not work, it is better to know early rather than later (say, after all the UI, database, etc have been designed).
It's like re-factoring, but not re-factoring. It's re-engineering. You can't engineer a business process, because it's business and not engineering. But you can re-engineer a business process (I guess by applying engineering processes to a business procedure).
I consider this word a "rubbish word thrown-in to confuse people".