Companies and organisations UK-wide who have sensibly invested in outsourced data governance are now enjoying working daily with error-free databases for the very first time. Their new database model is so expertly crafted; the problems previously created by conflicting data are no longer an issue. (Their new database system is designed to allow for only a single recording of each piece of data. Simple!)
Needless redundancy – A witch’s creation?
According to Ten things to avoid in a data model, a white paper written by Dr. Michael Blaha, (consultant and trainer in conceiving, architecting, modelling, designing, and tuning databases for dozens of organisations world-wide since 1994), avoiding needless redundancy – where a data item is recorded more than once – is essential.
Needless redundancy truly is the curse of so many information systems; whoever thought double or even multi-recording of the same piece of data in a database was a ‘clever idea’ should be burnt at the stake! Well, you get the gist…
By just sticking to the ‘single data piece only’ rule, the hassle of: error vulnerability, additional code-writing, latency updating, and more can all be avoided. Hallelujah.
The other nine things every data model should avoid?
These include:
1. Vague Purpose – The data model that doesn’t actually know what it’s for (imagine that)
2. Literal Modelling – Unquestioningly designing the model the customer describes. Hmm.
3. Large Size – Go for 200 tables max. (It’s a database, not a Sicilian wedding).
4. Speculative Content – Ahh, the ‘squirrel effect’ (data that’ll “be handy later”, but never is).
5. Lack of Clarity – If it’s ‘cryptic’ you’re after, just do a crossword. Really.
6. Reckless Violation of Normal Forms – A ‘shortcut’ to increased performance? No.
7. Parallel Attributes – The hallmark of a poor model.
8. Anonymous Fields – What’s in a name? EVERYTHING. Clearly described data is best.
9. Symmetric Relationships – Confusing, bamboozling, downright discombobulating!
Interesting stuff from Dr. Blaha indeed. But, at the end of the day, the key to great data modelling is placing engagement over formality. No doubt. How to do this? By working collaboratively and intensely with your customers, if you would like help with your data governance feel free to drop us a line.
Possibly Related Posts:
