CodingForums.com - View Single Post - Tagging system set up

Old Pedant · 02-08-2013, 10:13 PM

You know, a lot of this depends on how confident you are that you will *NEVER* want to do anything more than you are planning as of now.

I tend to avoid locking myself into a design just because my experience has always been that the customer (or my boss...same thing, really) comes back 3 months (or 3 years) later and says, "You know, I really want to also do ..."

And if I designed the DB right in the first place, I can say "trivial, let me just change this parameter." But if I did it wrong, then typically it means making changes to the entire DB schema. Which, even if it's not hard, can lead to hard-to-find errors weeks or months after then changes.

You seem so supremely confident that your specifications will *NEVER* change that I would say it doesn't much matter which way you go. I do not think you will see a significant performance difference in this particular case between the fully normalized and the somewhat de-normalized versions. But the only way to really know is run benchmarks.

02-08-2013, 10:13 PM	PM User \| #4
Old Pedant Supreme Master coder! Join Date: Feb 2009 Posts: 23,210 Thanks: 59 Thanked 3,996 Times in 3,965 Posts	You know, a lot of this depends on how confident you are that you will NEVER want to do anything more than you are planning as of now. I tend to avoid locking myself into a design just because my experience has always been that the customer (or my boss...same thing, really) comes back 3 months (or 3 years) later and says, "You know, I really want to also do ..." And if I designed the DB right in the first place, I can say "trivial, let me just change this parameter." But if I did it wrong, then typically it means making changes to the entire DB schema. Which, even if it's not hard, can lead to hard-to-find errors weeks or months after then changes. You seem so supremely confident that your specifications will NEVER change that I would say it doesn't much matter which way you go. I do not think you will see a significant performance difference in this particular case between the fully normalized and the somewhat de-normalized versions. But the only way to really know is run benchmarks. __________________ An optimist sees the glass as half full. A pessimist sees the glass as half empty. A realist drinks it no matter how much there is.