Applications in a way are like fish:

Expensive, cheap or look bad.

There is an interesting side topic about information being harder to make sense of as it gets older (difficult to stretch the analogy for this, so I won’t even try!). the more removed data gets from the processes and applications that created it, the less context you have. Particularly since those pesky applications and proceses are constantly changing and “absorbing shock”. Over time, new data does not line up comparatively with old data, making it difficult to understand in context. This is where metadata comes into play - a key to preserving context over time.

Good thought provoking post. You bring up a good point about the temporal nature of information and data. I think it is important that we don’t rush in and delete data as it could one day give us some insight as to what happened on a particular day in the life of that object, person, place, etc. I am accused of being a packrat, but I have proven over and over again that data is precious and we should always find ways to preserve it and make as much of it as possible available to us.

I think it is easier to find a place to put the data as opposed to finding data to delete. Every bit tells a story, and besides five minutes after you delete the file you will want it back, kind of a Murphy’s Law thing.

Cheers,
Tom West
Blog: http://www.ttoolboxes.ca/blog

Process is where data can be captured, application is where logic is applied to the data.

Processes changes, and they should - and at that stage data can be raw just reflecting the real facts.

The application of logic onto the data does not only change all the time, it should be allowed to be different for each user. That is why I whip out my HP 12 C every time I see some annual accounts!

Rip apart process and application - “applications” should be mere “templates” of logic, customisable of course.

Keep data in raw form, only apply (any) logic when needed - don’t mess up for the others!

(Your “middleware” clients may not like that idea though, making a living from extracting some usefulness from manipulated data as they are…)

MDM is important because the companies you cite are all trying to tie data together an…dare I say…normalise it. OK - wrong expression but basically they need get rid of the redundancy across multiple applications and instances of those applications. The perceived wisdom is to get it sorted out using meta data as the method.

ETL and DW is important for static data but modern thinking knows it’s needed as the bse for data in motion.

Come on - you know why Oracle is suing SAP. It woke up to the the fact TN has taken not just monkey nuts and it rests its business model for PSFT/SEBL/JDEC acquisitions on those 22% maintenance fees. Nowt to do with data.

Of course, information wants to be free. So does livestock. That’s why Joseph Glidden’s invention of barbed wire transformed the west. It made it possible for the owners of livestock to protect their property rights.

So let’s look at another aphorism, one I’ve been preaching since the 1980s, and maybe it will catch on pretty soon.

Hardware is free.
Software is expensive.
Data is priceless.
Customers are everything.

As Paul and Art sang, “Preserve your memories, they’re all that’s left you.” Teenagers want to be free, too, but if you aren’t careful about what you’re doing, instead of enjoying your “empty nest” years, you’re going to spend them raising your grandchildren.