>>>>Since the checksum is just digits, which are the same in all codepages, you couldn't have added any content which would be in a codepage different from the rest of the file.
>>>
>>>Not necessarily true.
>>>
>>>Only true if
>>>(1) txt file with UTF8 BOM, or
>>>(2) txt file without BOM
>>
>>Nope
>>
>>The chars added will follow the BOM / not BOM.
>>If you open up a dbf, add something and write, it will follow the rules of the dbf. It might corrupt the file - but it will be interpreted as a part of a dbf.
>
>Sure, the chars added will be after the BOM.
>
>And still, if you add chars ( = bytes in this case) below 0x80 and the text file has a UTF16 BOM, you are in trouble - guaranteed

I agree with this. But for the prog reading the file the bytes follow the UTF16 BOM. So they ARE in UTF 16 BOM - wether or not they are meaningless. :)

So you could not add something different to the underlaying format, because you can not change the definition.

Well, this is just text file. :)