Monday, 9 November 2015

Data Driven or Application Driven Development?

I'm the kind of person who believes that Data is really important and needs to be treated and dealt with properly. Of course we also need Applications that use that Data - one without the other is no use at all. But a battle I often have with clients now is that Data needs to be modelled and designed at the same stage as the Application is. Ignore the data model and database design and you will likely end up with a mess that will only come back to haunt you later on. (More on this in other blog posts to come).

I've just read a reasonably interesting blog post from Pythian on being Data Driven where this distinction is mentioned explicitly. At one point one of the panel mentions how some people continue to ignore the importance of Data and its design (my emphasis within this quote):
I recently I talked with someone who felt extremely behind the curve and they only recently did things like looking at data design as integral to the design phase of their system development. They had always, of course, reviewed design of the applications before implementing them, but data was always an afterthought. Applications create data, so what?

Now that they’re trying to get better at data, they had to actually explicitly make sure you got a data person in there designing the data, looking at the impact of what this application’s doing with the landscape or the information architecture they have and reviewing that before going off and actually writing a code.

That was new for them.
This is the kind of thing I have been on about for years - the Data is as important as the Application and needs to be designed properly up front. In fact, you can argue that the Data is more important in the long run. Why? Because Applications come and go but Data lives forever. Think about it - you often replace one Application suite with another, but you will almost always migrate over your existing Data i.e. you might throw away the Application software but you keep the Data itself.

And yet I continue to see software development organisations adopt Object Oriented and Agile Development methodologies wholesale and treat the Data as an "afterthought". Why, when the Data is at least as important as the Application?

No comments: