quote:

---

Yes much and much. There must be an interpreter behind it, taking all the strings, parsing them into tokens, making calls to OS subroutines, sending messages, waiting for the result, etc. If you guys have taken cources on computer language/compiler design must know that.

---

quote:

---

I think what Dan's trying to say, is that on today's typical server machine you're not going to notice a difference between Method A for creating a query and Method B. Now on a benchmark test, there would be a few milliseconds of difference, but to the normal viewer, they won't know any better, and that's what matters

---

I second that by part. And let this old guy tell you something, as a man who's first PC was a 8086 based machine (after Commodore 64), and had to implement programs under 640KB limit, a lot of firmware which had to fit in 1 KB and had to run in real time limits to control machines. I'll try to keep it short .

The problem is "the cost". In software engineering economics each step in the software cycle has its own costs (starting from requirements definition, through design, prototyping, implementation, test stages upto documentation and maintanence steps). There are a lot of related issues which are off topic here (commenting, versioning, use of CASE tools, etc).

But, one critical decision is the use of methodology. Depending on the application in hand one may implement the software using assembly language or Visual Basic. Why does a person choose assembly language? Because its incredible fast ! But the cost pf implementation AND maintanence is too high!

Most of the SW companies use Object Oriented coding environments and thus wasting resources of HW (Memory, processing time etc). This is also the case for MS$. As you know the requirements of each OS they publish doubles. But others will pay for them! This methodology will be OK if used on client machines, but on servers things will be different, as each connection will use the same machine.

Same applies here. Just to visualize it: If one page wastes 10 msec on each page, and there are 10,000 page views/day for this website, and there are 100 websites running on this machine... The total loss will be (0.01sec * 10,000 *. 100 = 10,000 sec) 2.7 hours of server processing time. Same applies for memory, connections, HDD access, bandwidth of any channel etc. Thus if you make a program that will run 10,000 times a day (even not in a loop :) you have to take care of these resources. Thats the "art of programming", thats the challenge and nice part of it.

On the other hand, there are other costs: cost of implementation, cost of testing, debugging, maintanence etc. If your webpage design is highy changing in time (anywhere inbetween static and dynamic website) you will not be able to use web design tools if you use response.written html. You have to look at the code and and see how large an image is and design another one etc.

There is a trade-off and you have to decide for the best cost.

Thank you for your time.

Think Pink