Rick,

I agree with all your points, been writing code long enough to know real world business apps are not doing one million loops of calculating a sin() function :). My post was not about comparing VFP speed for real-world apps but just to get a measure of how far behind the curve VFP has fallen in brute force processing, if indeed it has. Obviously VFP will not compare to compiled languages - we even still use an assembler DLL to perform a string manipulation function which performs 1000x faster than the VFP equivalent :) I am interested to read that VFP will still compare favorably with scripting languages.


>Jos,
>
>Here VFP is probably going to fare a lot worse than any native compiled language because you're dealing with low level processor directives. Any compiled language like C,C++,Pascal etc is going to be much faster at this. .NET languages which eventually compile down to machine code as well should be much faster although not as drastically because of the intermediary JIT compilation. I'd expect Java to give results similar to .NET's as well.
>
>Pure scripting languages like PHP and even Ruby probably will be on par or slower than VFP.
>
>Realistically though code like this doesn't matter much because how often do you write code like this? Code that matters is typical CRUD data access and working against some sort of higher level output API (expecially for Web apps) which are most likely implemented in machine code by the fraemwork. Language performance is less and less important - it's the platform that matters. This is especially true for Web frameworks where the core code operations in the framework tend to be native code that is called by scripting languages.
>
>VFP is plenty of fast for just about any kind of app IMHO. Where it lacks is in scalability as it's designed to run as if it was the only thing on the system (even when running as MTDLLs).
>
>+++ Rick ---
>
>
>>Hi All,
>>
>>I am doing a raw speed test on some very simple operations to compare basic VFP operations with another language. Would some of you care to do a similar test against another language of your choice? e.g. Python, PHP, VB.Net, whatever. I am looking for VFP speed vs. language_of_your_choice speed on the same machine. Here are two simple tests manipulating some numbers and building a string in VFP
>>
>>
>>

>>CLEAR
>>tStartTime = SECONDS()
>>ta=0
>>FOR i = 1 TO 1000000
>>  ta = Log(i) + Cos(i) + Sin(i) + Log(i) + exp(i)
>>next  
>>? SECONDS() - tStartTime 
>>
>>*
>>
>>tStartTime = SECONDS()
>>ta = ""
>>FOR i = 1 TO 1000000
>>  ta = ta + "xx"
>>next  
>>? SECONDS() - tStartTime 
>>