I guess Kevin Lawrence did a good job testing. But when I see these test, I'm a little skeptical about the results. If he redo the same test again tomorrow, he's likely to get a different result.

What I get out of these test is that, even machines have good days and bad days. I think you can even say that they have good moments and bad moments during the day. So, a query that took just .005 seconds just a few seconds ago can take a wopping 0.1 second next time around.

Personally I don't care about these tests. As long as the tool returns the result in resasonable amount of time. You can only do so much with testing.

Then again, if I don't care about these tests, why am I responding to this thread? ;)


>Kevin;
>
>I have to thank you for your effort and interest with this subject, as it elicits comments, thoughts, and makes the attempt to compare methodologies. It is reminiscent of George Goley’s work with FoxPro, determining what worked fastest, etc. Also thank you for the code examples.
>
>One question – you say you are comparing Visual FoxPro and C#? Would it be more accurate to say you are comparing Visual FoxPro’s data engine and ADO.NET? I am not trying to take away from your work or split hairs, and apologize if it seems I have done so with such a queston. Perhaps it is just my engineering background that makes me ask such questions! :)
>
>Another question that comes to mind is how about functionally (speed) using VB.NET vs C#? That is another issue I know but one idea spawns another!
>
>Comparison with SQL Server and other back ends would be of interest. Perhaps others can add to work such as yours. My personal preference is to use SQL Server with stored procedures.
>
>If we defined ten scenarios of application languages and back ends, with ten different developers, we might have at least 100 different results!
>
>In the end I do not think there should be much difference in the results. It is most important to get the data set quickly enough so the user does not complain!
>
>I have always been interested in “speed of data retrieval” and have many tools I have developed for that purpose, going back to dBase II. It becomes academic at one point, and soon you choose which way to go based upon your work.
>
>Now if I only had a Cray computer!!!!! My wife said they would never build a computer fast enough for me! Well, a Cray might just do the trick! :)
>
>
>
>Tom