A few points to make: One of the first things to remember when making claims such as this is "propogation of effects." While a choice may appear to be the more financially beneficial one at first, the effects of that choice may very well end up costing you lots and lots of money -- and that's definitely not a good thing (I think we can all agree on that.)
For example,
Don't use Oracle...
Okay, so it may initally cost effective not to use Oracle because the help is more expensive and the license costs an arm and a leg, but you have to think about what happens when you launch your website. IMHO, Oracle running on a Sun or similar *NIX box is inherently more stable and will provide better performance than MS SQL on an NT box. If my website is/.'ed the first day it goes live, survives the onslaught (where MS SQL would choke and die,) I think that extra $$$ is well worth it. I think the shareholders would agree as well.
Don't use Unix (any flavor)...
Again - think about it. IMHO, a Unix-based box will crash less and handle a higher load than a similarly equipped MS SQL box (sure -- if you give MS SQL twenty clustered 4-way Xeon boxes it can handle a high load, but there goes your next years' profit out the window when you need to buy a room at Exodus and hire 5 full-time admins to keep it alive - by clicking checkboxes, of course)
Don't support Linux (yet)....
I'm not going to touch this with a 20-foot pole....:)
In the end, the cost of your consultants and the initial fixed cost of the software/hardware itself can nicely be offset by the amount of money you'll save in the future by having a fast, stable website that doesn't crash and makes the users happy. And what does that get you (assuming you have an idea that doesn't suck?) -- $$$.
Slashdot Mirror
A few points to make: One of the first things to remember when making claims such as this is "propogation of effects." While a choice may appear to be the more financially beneficial one at first, the effects of that choice may very well end up costing you lots and lots of money -- and that's definitely not a good thing (I think we can all agree on that.)
For example,
Don't use Oracle...
Okay, so it may initally cost effective not to use Oracle because the help is more expensive and the license costs an arm and a leg, but you have to think about what happens when you launch your website. IMHO, Oracle running on a Sun or similar *NIX box is inherently more stable and will provide better performance than MS SQL on an NT box. If my website is /.'ed the first day it goes live, survives the onslaught (where MS SQL would choke and die,) I think that extra $$$ is well worth it. I think the shareholders would agree as well.
Don't use Unix (any flavor)...
Again - think about it. IMHO, a Unix-based box will crash less and handle a higher load than a similarly equipped MS SQL box (sure -- if you give MS SQL twenty clustered 4-way Xeon boxes it can handle a high load, but there goes your next years' profit out the window when you need to buy a room at Exodus and hire 5 full-time admins to keep it alive - by clicking checkboxes, of course)
Don't support Linux (yet)....
I'm not going to touch this with a 20-foot pole....:)
In the end, the cost of your consultants and the initial fixed cost of the software/hardware itself can nicely be offset by the amount of money you'll save in the future by having a fast, stable website that doesn't crash and makes the users happy. And what does that get you (assuming you have an idea that doesn't suck?) -- $$$.