Only one improves with age. With apologies to the originator of the phrase – “Hardware is like fish, operating systems are like wine.”
What am I talking about? Processes change, and need to change. Baking data into the application is a bad idea because the data can’t then be extended in useful, and “unexpected ways”. But not expecting corporate data to be used in new ways is kind of like not expecting the Spanish Inquisition. But… “NOBODY expects the Spanish Inquisition! Amongst our weaponry are such diverse elements as: fear, surprise, ruthless efficiency, an almost fanatical devotion to the Pope.” (sounds like Enterprise Architecture to me).
I didn’t go quiet Dennis, I just didn’t post this. So if enterprise apps are so cool- why are SAP, IBM and Oracle all competing in the Master Data Management Space? If enterprise apps cut the mustard why is ETL and datawarehousing still a strong market. If the app matters more than the tax algorithms why is Oracle bothering to sue SAP?
I actually really don’t like the bracketing of data extensibility as The Intel Inside. In fact I hate that way of articulating it. Because to my mind Intel may be a standard, but its not an open standard (well it wasn’t until AMD started to really nail it).
Of course process matters but its better to crucial to invest in having the right data. Arguably the pitch SAP traditionally made was not that it would sell you a suite of apps, but rather than it would bring all your corporate data into one place so that you can make business decisions based on facts rather than suppositions. SAP’s story today still talks clearly to this corporate data standardisation.
Information wants to be free. the Intel Inside slogan came from a monolith/island era. That’s not where we’re going next.
In this industry we tend to focus on the need to separate data from presentation, but of course we also try and enforce a separation beween data and business logic. Why? Because business logic quickly begins to smell. From a pace layering perspective.. oh wait, y’all probably don’t know what pace layering is. Its this cool idea from Stewart Brand, explained here (by Donna Maurer)
“Complex systems can be decomposed into multiple layers, where the layers change at different rates. The “fast layers” learn, absorb shocks and get attention; the “slow layers” remember, constrain and have power. One of the implications of this model is that information architects can do what they have always done–slow, deep, rich work; while tagging can spin madly on the surface. If it is worth keeping, it will seep down into the lower layers.”
Which talks pretty well to my point. Has anyone formally mapped Model-View-Controller to pace layering? I doubt it. But if we look at business architectures data has the longest shelf life, followed by business logic, followed by presentation. Sometimes the data stays the same but the data model changes, another layer. Presentation goes off even faster than fish [ed: what does go off faster than fish, or is this metaphor b0rked? Why do shops keep investing in mainframes? Because they are investing in data. Why do people keep buying apps? Because their old ones went off.
Just to abuse the metaphor a little bit more there are some interesting things to note about wine. Not all wine is designed to be kept. Not all data increases in value over time. If organisations spent more time working out what wine was worth keeping, and which had turned to vinegar, they would be a lot better off. Retain the data that has value. In some applications- market risk for example-the only reason to keep historical data is regulatory requirements. Operational risk on the other hand needs to more cognizant of trending. For market data risk freshness matters.
“To me, data is valuable for understanding what is going on inside a process. The data may suggest an alternative course of action. But that in itself implies the application of process steps and will have been the result of pre-presentation filtering and refining, via business processes.”
. The process provides a context for the data, sure, but could the process or application be different? Regardless of whether or not corporate data has a life outside the process, which of course i think it does, it would still pay organisations to really think through their data, and what its for. We don’t want it all in the cellar, after all. Sometimes you need a drink. Anyway- what do you all think?