Tore --

Depends what you're multiplying by two, doesn't it?

Go back to my example. Suppose I have a number of algorithms I'm investigating. I'm doing so specifically because they are "slow" and need improvement. Some are slower than others. Some are unbelievable slow -- in fact, they are ten times slower, as they take ten times as much time to complete.

I maintain that this is a matter of perspective. I believe that if you interpret "slow" as meaning taking a particular interval of time, then "ten times slower" means taking ten times that interval of time.

How would you answer the question "how much slower is one?" I think it's valid in this context to say "it's ten times slower" (clearly, to me at least, meaning it took ten times as long).

Cheers,

Jim


>Hi James,
>
>No, it doesn't. Please bear with me if I may use some wrong terms in this explanation.
>
>To put it short, if '10 times slower' should be correct, then what about 'ten times faster'? If you drive 30 MPH, and your friend drives the double speed (2 times=double), he will drive 60 MPH. Adding the word 'slow' will be technically meaningless because 2x30 MPH = 60 MPH no matter how you bend and twist.
>
>
>>... and pardon me for jumping in as well ...
>>
>>But if I am evaluating a number of algorithms to solve a problem, all of them known to be "slow", doesn't "10 times slower" mean it takes 10 times as long? (If you understand "slow" to be defined as "taking a given period of time", then "10 times slower" sounds like a perfectly reasonable construction to me.)
>>
>>Jim
>>
>>
>>>Hi Tamar,
>>>
>>>please pardon me for jumping in.
>>>
>>>I want to arrest you for using wrong words. You have arrested other for using words such as recurse, so you have put yourself in a glass house, as we say in norwegian.
>>>
>>>Nothing can be 10 times slower, that's impossible, but it can take 10 times the time to execute. What you mean to say is that it is one tenth of the speed. I have many years of advanced math from high school, and my teacher would have turned over in his grave if he heard '10 times slower'. When you multiply any positive value by a value higher than one, you always get a higher value. So technically '10 times' and 'slower' is a contradiction, or at lease they don't belong together. Just like '10 times less' or '10 times smaller' also are samples of wrong use of words.
>>>
>>>I hope I have made myself clear, and that you understand my explanation?
>>>
>>>NB! No pun intended. :-)
>>>
>>>>>RE: Major Performance Penalty
>>>>>
>>>>>Is this type of thing still an issue? I mean, CPUs are so fast now ... not that performance isn't an issue (I have one calculation that runs for 20 minutes), but rather that it seems that the overall design seems to be of much more major of a concern that the internals of how VFP does things.
>>>>
>>>>Just looked at my notes on this (which happen to be in the Hacker's Guide, where everyone can read them <g>). At my last testing, having:
>>>>
>>>>

>>>>*MyProg.PRG
>>>>PROCEDURE MyProg
>>>>* Code starts here
>>>>

>>>>
>>>>was 10 times slower than:
>>>>
>>>>

>>>>*MyProg.PRG
>>>>* Code starts here
>>>>

>>>>
>>>>in VFP 6 and earlier, but the penalty was gone in VFP 6 SP3 and VFP 7. Haven't tested since then.
>>>>
>>>>You also raise a good question about the importance of such things in an era of fast CPUs. It depends, of course, on how many times you make such calls.
>>>>
>>>>Tamar