Re: [xml-dev] Argument: Software design is important, data design is not

On Tue, Nov 12, 2013 at 9:32 AM, Costello, Roger L. <costello@mitre.org> wrote:

Software is king!

But as in most absolute monarchies, the taxes are high and the benefits doubtful.

Data design is only needed to the extent that you document and define what it contains; other than that, do whatever you want. Software design is important, data design is not.

What is data design anyway? A typical response is: a good data design is one that enables software to do its job better/easier. Again we see that data is just in a supporting role to software; data is not the lead character (to use an acting metaphor). If the data is to be consumed by diverse software packages, a single data design cannot make all the software packages do their job better/easier, so why bother doing any design?

Data designs like the hierarchical file system are so powerful and useful that nobody would even think of scrapping them and going back to purely flat disk directories (or for that matter, a system in which all programs have their own individual tracks on the disk which they use exactly as they like). We pay the cost of design, ideally, just when the costs of non-coordination outweigh the advantages of non-coordination.

Arguing that one data design is better than another is a waste of time and, in fact, it's meaningless. Spend minimal time on data design and create great software designs.
---------------------------------------------

What's wrong with that argument?

It's shortsighted. It assumes that data is ephemeral, but in general data outlives software. The arrogance of kings arises out of their belief that their people were made for them, and not they for their people.

GMail doesn't have rotating .sigs, but you can see mine at http://www.ccil.org/~cowan/signatures