>>>>>But, what about when open vfp, issue a DO FORM on the problem form, then issue CLOSE DATA ALL, and then try a MODI FORM on it? In other words, I only want to know if caching is affecting anything or not.
>>>>
>>>>>
>>>>>This is important to my testing & diagnosis attempt.
>>>>
>>>>In this case MODIFY FORM is slow.
>>>>Caching does not seem to affect this.
>>>>The data files MUST BE OPEN when you issue MODIFY FORM.
>>>
>>>Good, thanks. I'll do a little more testing tomorrow, and if I find anything else, I'll make a post about it.
>>
>>I'm still not sure what I'm reading here.
>>
>>If the data files must be opened to get reasonable performance, doesn't this mean that cache **IS** implied????
>>
>
>
>Hard to believe you say this! What part was not clear? If the line "Cashing does not seem to affect this." confuses things, just ignore it, I don't think it makes a difference.
>I don't want to change the subject to caching, although it is a big source of frustration for me. I have no idea what is, and what is not cashed, and I believe that nobody, especially from MS, will give a straight answer (see my recent attempts tofind out how EXE files lose track of class lib that are included in the project.)
>
>
>>For what it's worth - on a standalone workstation - I can do 5,000 SEEKs to each of 3 tables in an elapsed time of 135 seconds, and when that finishes issue a CLOSE TABLES ALL and a SYS(1104) and a FLUSH and then run the app immediately again and it will run in 1.31 seconds!!!
>>
>>I know that things are different when tables are network resident, but still get one wondering about the 'effect' of CLOSE TABLES ALL and SYS(1104).
>
>In "my case" (see the subject title) this does not apply. As I said, as soon as you close the tables MODIFY FORM is as slow as it ever was. Even if they may be cashed somewhere, MODIFY FORM does not seem to care.

Whoa, Doru, no need to get your nickers in a knot, ESPECIALLY when there seems to be some question as to whether one correctly understands "caching".

In another reply (to BruceC, I think) I said that I felt that caching was involved, but something ELSE TOO.

Now, more than ever, I feel so! Am I wrong in having read YOU say that as long as the tables were opened by some process on the same workstation - even FPD/FPW - the MODI FORM was fast??
Believing that that IS what you said, then guess what - opened by another process of the workstation means CACHED!

Now just what it is in MODI FORM that works the way it is observed is what I'd like (and you too) the answer to.
>
>>
>>cheers