There is more than one way to skin a cat
Americans sometimes use a funny little
expression: There's more than one way to skin a cat.
This is a (kind of creepy) way of saying that there are several different ways to accomplish a given task. And this is my challenge. My developer side says to do it one way, but my admin side might have a different perspective. In the end, a decision must be made one way or the other. Case in point:
Recently, I completed the first part of a source code control and build process management implementation (FYI, Teamstudio knows quite a bit about source code control). The team included Chris Blatnick of Interface Matters fame, Kevin Petitt of SuperNTF fame, and Jamie Magee of Martin Scott and NoteMan Tools fame. It really was special for me to be able to teach these guys anything.
All of us are familiar with other tools, as well, like the ones from Ytria. The point here is that you can be so much more productive if you invest in some tools. Of course, we all prefer that you choose *our* tools, but what we really want is to help you deliver the best solutions for your users in the most cost effective way possible. At the risk of repeating myself, that generally means you need to invest in some tools. (For what it's worth, I definitely like the 'free' ones, but you should always keep in mind that you get what you pay for).
Anyway, here's the challenge. I came across an agent that created a document collection by doing a search. My first reaction (from the developer side of my brain) was to just create and use a view. Further review showed that the agent ran once a week and only took 20 seconds. That's when my admin side took over, and said: Why have the server maintain the index for this view all week long, when it would only be used once a week?
Any thoughts? Is a search always bad? How much overhead is noticeable when maintaining the view? Multiply this by dozens of databases. How would you decide?
This is a (kind of creepy) way of saying that there are several different ways to accomplish a given task. And this is my challenge. My developer side says to do it one way, but my admin side might have a different perspective. In the end, a decision must be made one way or the other. Case in point:
Recently, I completed the first part of a source code control and build process management implementation (FYI, Teamstudio knows quite a bit about source code control). The team included Chris Blatnick of Interface Matters fame, Kevin Petitt of SuperNTF fame, and Jamie Magee of Martin Scott and NoteMan Tools fame. It really was special for me to be able to teach these guys anything.
All of us are familiar with other tools, as well, like the ones from Ytria. The point here is that you can be so much more productive if you invest in some tools. Of course, we all prefer that you choose *our* tools, but what we really want is to help you deliver the best solutions for your users in the most cost effective way possible. At the risk of repeating myself, that generally means you need to invest in some tools. (For what it's worth, I definitely like the 'free' ones, but you should always keep in mind that you get what you pay for).
Anyway, here's the challenge. I came across an agent that created a document collection by doing a search. My first reaction (from the developer side of my brain) was to just create and use a view. Further review showed that the agent ran once a week and only took 20 seconds. That's when my admin side took over, and said: Why have the server maintain the index for this view all week long, when it would only be used once a week?
Any thoughts? Is a search always bad? How much overhead is noticeable when maintaining the view? Multiply this by dozens of databases. How would you decide?
Category
| Permalink |
|
|
|
|
|
Comments
Posted by Toby Samples At 03:47:34 PM On 02/02/2012 | - Website - |
In general I prefer the search as it provides a lot more flexibility. I have seen so many databases with lots of views where nobody has any idea which views are used and which not. It's hard to keep track of these views and maintaining these costs a lot of performance.
Cheers - Michael
PS: Yes, I have been to A. Pollack's session about Performance Programming... ;-)
Posted by Michael Ruhnau At 07:58:10 AM On 02/03/2012 | - Website - |
Posted by John Kingsley At 11:15:11 AM On 02/03/2012 | - Website - |
Posted by florida gator forum At 01:23:54 PM On 04/27/2012 | - Website - |
Posted by Coach Outlet At 10:00:43 PM On 05/17/2012 | - Website - |
Posted by Coach Outlet At 10:01:05 PM On 05/17/2012 | - Website - |