This is the problem.
Numerics in Microsoft basic (GWBASIC, QBASIC, BASIC COMPILER 2.0, and others basic) may not be IEEE.

Thanks Fabio.

>>No success.
>>
>>In RANDOM basic file there is a value of CHR(0xD4)+CHR(0x08)+CHR(0x03)+CHR(0x85) for a price, that it sees in screen as 16.37931061
>>
>>CTOBIN() can't help me.
>>Thanks.
>>
>
>Then it is not a IEEE single-precision format.
>
>the single-precision format for 16.37931061 it is
>

>CHR(0xD4)+CHR(0x08)+CHR(0x83)+CHR(0x41)
>

>
>
>>>>I need read ramdom Microsoft basic files, with binary numeric fields.
>>>>I need some VFP code routine that emulate the CVS(), CVD() basic functions.
>>>>
>>>>Thanks.
>>>
>>>
>>>On VFP9:
>>>

>>>*CVD()
>>>* correct
>>>CTOBIN(8bytes,"BRS")
>>>* uses
>>>CTOBIN(8bytes,"BS")
>>>

>>>
>>>I hope this bug will be fixed on SP2.
>>>
>>>

>>>* CVS and CVD
>>>CTOBIN(4or8bytes,"NRS")
>>>

>>>
>>>Remember, VFP doesn't support 32bit floating point (single),
>>>then the conversion to double can add a rounding error.
>>>
>>>example
>>>

>>>* double
>>>? CTOBIN(0H+BINTOC(0.1E0,"BRS"),"NRS")
>>>
>>>* single
>>>? CTOBIN(0H+BINTOC(0.1E0,"fRS"),"NRS")
>>>