>>>>>>If the routines of AT and OCCURS they were both written best,
>>>>>>time for execute the request would have to be practically identical.
>>>>>>
>>>>>>Then, exist native VFP command that C++ can speedup.
>>>>>
>>>>>Ok, Fabio, when will you be submitting your resume to MS to be a SDE on the Fox team, since
>>>>>you obviously know better than anyone who currently works there?
>>>>
>>>>Crass response (yours)!
>>>
>>>True. I get tired of listening to the same thing over and over, though.
>>
>>I for one value the contribution. Given that you are a PUTM why not update your twit filter?
>
>Because I value his contributions, too. I particularly "like" that LOCATE/NULL bug he found yesterday.
>
>I just don't consider his comments on the competence of the VFP programmers to be "contributions". :-)

I don't see any "comments on the competence of the VFP team" per se. I do see that Fabio has spotted an enormous difference in performance between two functions that ought to have similar performance and this is especially the case where the position of the desired instance from AT() is the same as the number of instances (this assumes an even distribution of the target string in the source string). The observation is valid even if the delivery is terse.

Given that we techies are renowned for our lack of inter-personal skills and that Fabio’s first language is obviously not English, I think your response is both unwarranted and now that you’ve clarified what pressed your button, erroneous.