T-SQL Tuesday #100

When T-SQL Tuesday started back in late 2009, I have to admit that I wasn’t sure why Adam used three digits. I didn’t know whether it would reach 10, let alone 100. 2018 seemed so far away – if you’d asked me if I thought T-SQL Tuesday would still be going in 2018, I wouldn’t’ve had a clue. And I definitely wouldn’t’ve thought I’d still be contributing, having never missed one. I’m pleased I haven’t missed any though, because it keeps me writing. I enjoy having a writing assignment each month – it forces my hand, even though it’s self-inflicted.

Anyhow…

In his invitation post, our host and T-SQL Tuesday owner Adam Machanic (@adammachanic) makes some interesting points about the types of topics that have made up the list so far, that there has been a significantly high proportion of topics that are not technical in nature, showing that there is a large interest in personal development as well as technical. I suspect this implies that people want to know what people are thinking, not just how to solve problems. People can ask for technical help on a variety of sites, but perhaps T-SQL Tuesday is unique in its ability to ask people for their opinions. I do think that the next hundred topics will be less technical and more personal.

Adam asks what things will be like after another hundred months’ time, and it’s an excellent question.

I think people will continue to become more and more interested in their personal development rather than their technical development. And I think this will increasingly become the shift of the world. Many years ago, we saw robots replace factory lines, and today we see self-driving cars. As time continues, I’m confident that this trend will continue, and that “regular folk” like us will need to solve fewer and fewer technical problems. Performance tuning seem to be already on the way out, with some of the advancements that we’re seeing in Azure SQL Database. AI advancements allow us to create systems that are smarter than ever before, and database consultants are becoming business consultants.

And this is because I see ‘business consultancy’ as being one of those ‘hard’ tasks. In school we learned about P/NP problems – P being the set of problems that were solvable in useful (the real word is ‘polynomial’ time, but I figure that the real thing is whether it can be solved quickly enough to be useful), and NP being the set of problems that weren’t, despite it being quick to demonstrate that a particular solution was a valid solution (this is just my paraphrase of it all – don’t beat me up for being exactly correct). Obviously you can just test every possible answer to see if it’s a valid solution, but that’s not quick enough to be useful. The world of computer science is still finding ways to turn problems from being NP into P, and we shouldn’t have to spend time thinking about P problems.

If I assume that consulting will always be needed, why isn’t it possible to turn it into a P problem? I suppose that’s like the question of whether a computer can ever produce beauty on a reliable basis. And I think it can. It should be possible to analyse a given situation, and produce a strong-enough plan of attack to produce sensible advice. A self-driving car has a very restricted set of parameters in which it can make decisions, but ten years ago we would’ve all considered this was too NP hard to become useful. Medical consulting has been replaceable (to a certain extent at least) by ‘expert’ systems for many years. So I suspect that business consulting and certainly data consulting ought to be able to go the same way.

…maybe not in the next 100 months though. And hopefully not until I’m ready to retire. I might still be writing a blog post every month, mind you, unless that can be automated as well.

@rob_farley