Back to some semblance of normality, able to attend several presentations without being preoccupied with having to do the same.
Attended an interesting talk on rule based applications in SQL. Some food for thought here, but once again the slides were stuffed full of text, and code text at that. It’s complex, and no way was the speaker going to show us how to write the appropriate scripts. The content of the presentation ought, given the tight time scales, to have been at one level of abstraction higher than was the case.
This was followed by a presentation on surrogate keys. Well, I have always been interested in design issues, and wanted to see how others tackled this subject. The advice here was to always use surrogate keys unless there were overwhelming reasons why one should not. Is this advice that I would disagree with? In the immortal words of Ian Richardson – ‘you might say so, but I could not possibly comment further’, so I won’t.
My next port of call was to try and learn a bit more about the Grid. Nicholas Donatone gave an excellent low-key presentation on the subject, using simple to understand examples, and involving the audience as he went. His slides were relevant summaries of the points he was making, and were pleasantly clear. No histrionics, just a comfortably delivered presentation. How nice!
The last port of call did provide histrionics. Scott Ambler, on agile database techniques. A good friend warned me beforehand that I should avoid this presentation if I didn’t want to be horrified. I needed to know what all the fuss about agile was, so I went. And I was horrified. Seriously. Full of provocative statements, like ‘you do not need to do data modelling up front, its a waste of time, its the kiss of death’. The classical database designer (in which group I would include myself) was repeatedly bad-mouthed, Scott having previously said that they and the new agile generation need to talk more. That’s a motivational start. The extraordinary thing was that he argued in favour of doing things that certainly I (and a couple of other members of the audience, judging by their input), have been doing for years and years. I was left wondering where the beef was… Actually, I would just love to organise a cross-table debate between Scott and one of our more conventional database gurus. I wonder…
If Peter Robson is horrified, there is something seriously, seriously wrong. I do hope you will expand on this topic when you have an opportunity.
It sounds like the current approach of “code first, gather requirements later, and always blame the database for performance problems” is not going away, but getting stronger. At least it will keep the performance specialists employed for many, many years to come.
Nice to hear from you again, Dan! I’ll elaborate on this matter when I get home. But you have summed it up in one – never mind the design, lets just start hacking (well, a small exageration perhaps).
I’m am looking forward to your full treatment of the presentation.
Sue Harper has an interesting blog entry and link to a presentation about Agile development that is a little more reassuring. (http://sueharper.blogspot.com/2006/03/agile-again.html)
This makes me think, “Good organizations will design/develop/implement good systems regardless of their tools & techniques; Poor organizations will design/develop/implement poor systems regardless of their tools & techniques”