Ok, so this is actually Sean writing this time.
I thought I’d hit the ground running with a great topic that came up just this morning.
The issue is how do you prioritize tasks in code? So let’s say you’re going to process a list of DBs for processing, but you also want to do them in a specific order. You’ve got 2 choices really when deciding on how to do this: Priority and Weight.
This one is probably the most common. This is say when you have a list of objects and you rank them from say 1-10. In this case 1 is the top priority with 10 being the least. This is how a priority system works. Items move up the list to the top position.
A weight system works just the opposite way. You give more weight to items by giving them a higher number. This extra weight makes them more important so they get processed first. So in the 1-10 list, 10 would be processed first because it has more weight.
So what’s the difference between these? Well a priority system I think is easier with fairly static priority code. You’re not going to be making any real changes to the priorities or the objects very often.
A weight system is better when your list and priorities are more dynamic.
Let’s take a look at an example:
You want to process some DBs for something. You pull in the list of DBs and you want them in a specific order. So you put them in #DBs with an Ordering col.
If there won’t be any new DBs in the list for a long time, and you’re not likely to change the order, then a priority system is good for you. However, if you’re going to be adding more DBs, and those DBs will possibly even be higher priority than the current ones, then a weight system is best. The reason is that if you’ve already got the number 1-10 populated, and something new pops up in the list that’s higher priority than anything else, you’d have to set that to 1 and then shift all the others down by 1. So what used to be 1 is now 2, what used to be 2 is now 3, etc. However, with a weight system, you can add as many new items as you like, and you just keep increasing the number. So in this case the new item would be assigned 11 because it’s more important than anything else and has more weight.
Of course if you have to arrange the weights of any of the lower ones it does become just a bit more tedious. In this case you’ll have possibly make another sub-group column. This way you can keep your major groups and then just change priorities or weights in the sub group. So let’s say you’ve got DBs grouped from 1-100 using either method, but we’ll say it’s by weight for now. If you have to make one DB a higher weight by one you’d have to shift everything else up one.
However, if you have that DB in a major group of say 10 DBs, and they all have weights within that group then you only have to change the weights within that one group. So whereas you would before have had to change the weights of 30 DBs, now you only have to change under 10.
I tend to use a weighted system more, but I have nothing against a priority system. And I prefer the 2 group system even more.