>>The other version still doesn't seem to always work. I switched it to use the C Standard Library's RAND() and SRAND() functions:
>>
>>

CLEAR
>>lcEnc = encrypt("Test string here")
>>? lcEnc
>>lcDec = decrypt(lcEnc)
>>? lcDec
>>
>>
>>
>>FUNCTION encrypt
>>LPARAMETERS tcString
>>LOCAL lcString
>>
>>    seedRand(0)
>>    lcString = SPACE(0)
>>    FOR lnI = 1 TO LEN(tcString)
>>        lnRand      = ROUND(getRand() * 255.0, 0)
>>        lcChar      = SUBSTR(tcString, lnI, 1)
>>        lcChar      = CHR((ASC(lcChar) + lnRand) % 255)
>>        lcString    = lcString + lcChar
>>    NEXT
>>    RETURN lcString
>>
>>
>>
>>
>>FUNCTION decrypt
>>LPARAMETERS tcString
>>LOCAL lcString
>>
>>    seedRand(0)
>>    lcString = SPACE(0)
>>    FOR lnI = 1 TO LEN(tcString)
>>        lnRand      = ROUND(getRand() * 255.0, 0)
>>        lcChar      = SUBSTR(tcString, lnI, 1)
>>        lcString    = lcString + CHR((ASC(lcChar) + 255 - lnRand) % 255)
>>    NEXT
>>    RETURN lcString
>>
>>
>>
>>
>>FUNCTION seedRand
>>LPARAMETERS tnSeed
>>
>>    DECLARE INTEGER srand IN msvcr71.dll as cSrand INTEGER nSeed
>>    DECLARE INTEGER rand  IN msvcr71.dll AS cRand
>>    cSrand(tnSeed)
>>
>>
>>
>>
>>FUNCTION getRand
>>LOCAL lnRand
>>
>>    RETURN cRand() / 32767.0
>>

>
>This code works every time. Thank you. But (don't take it as a criticism of your approach).
>The decripted string looks the same every time the original string is the same. So, as I mentioned in one of the previous messages, the downside is that a user can simply copy a value from one record to another.
>It would be nice if the encrypted string would look different even for the same unencrypted string.
>I may just have to live with this.
>Thank you.

Use the RECNO() as the seed.