Hi!

Well, hacking programming could improve performace a lot. I will not wonder if VFP is change-self app, when machine codes are replaced on the fly ;)) That is one of the approaches to improve app performance. I do not know other issues that can cause the problem with look ahead logic and caching logic of new CPU's. Maybe I should refresh my Assembly language and processors knowledge (did not used them already 5 years...)

>Hello Vlad,
>
>Another area which may contribute to such problems is newer CPU's can process commands out of sequence, with their own look ahead logic and buffering methods. Possibily there is some flaws in the hardware logic not accounted for in the C compiler used by the fox development team??
>
>Bob
>
>>Hi!
>>
>>It is the result of the old times technique to wait for something (processor ticks). Bad knowledge of the multi-threading and OOP/structured programming is also an issue. The VFP engine is developed for a lot of time and it seems it is not changed, just patched. That is why we have so many weird things happening.
>>
>>Another reson might be kust a huge amount of different functions that are hard to maintain. Saym MS required to fix all routines to use new memory management model insode of the data processing commands. There could be so many places for fixing that something is missed. When testing, missed thing was not observed. But in real life that thing pops up with weird behavior.
>>
>>I can bet MS did not ever tested VFP by accurate stress testing. Something like posted here could be found that way.
>>
>>>It seems to me that VFP is getting more and more of these "Going too fast for itself" glitches...
>>>I'm ass/uming that this is to do with Processors getting faster etc.
>>>Do you know if other development environments are getting caught in the same way?
>>>
>>>
>>>>Hi!
>>>>
>>>>Correct. That si why I told about black VFP box. We had similar fundamental functionality problems in VFP.
>>>>
>>>>>>You should use an integer as a primary key... you are taking too much of a chance that a user will not hit save at the same time. This is really an unnecessary risk.
>>>>>
>>>>>While I agree with all these responses about not using a date for a PK's
>>>>>Doesn't this thread strike anyone as a bit strange?
>>>>>
>>>>>Carol is basically doing a Scan loop, adding 1 to a variable and shoving that variable into a field and this is not working
>>>>>Pretty fundamental don't you think?
>>>>>
>>>>>Does this mean that VFP has trouble adding 1 to a Datetime value?
>>>>>
>>>>>Just my .02