VB Forum / Controls / July 2008

You certainly could create a class which handles arbitrary precision
math, or likely any other such thing you need.  I've seen implementations
like complex numbers, arbitrary precision using byte arrays, and some
other wild stuff.  No sample code to share, though - sorry.

Using just two Doubles might get you to within a few bits of
quad-precision.  I'm not sure if I would approach it quite that
simplistically, though.  For that sort of thing, Byte Arrays have
become my hammer of choice - sometimes you can emulate an external
storage type right in memory, and greatly reduce the translations
necessary for API calls or I/O using the target type.

Of course, in VB6 you can't overload operators, so you pretty much
won't be able to use those.

    Bob
    [MVP Classic VB]
--

Curious.  What do you mean VB6 or VB2008?  Are you saying you're indifferent
as to which you use?  In that's the case, you really first need to pick one
or the other. And then, depending on which you pick, post your question in
the appropriate newsgroup.  Also, examples of the kind of data you're
working with might have been helpful.

How many decimal places are you dealing with? If 4 or less, you might find
working with the Currency data type to be better than Double (at least in
VB6). The Currency data type is a scaled integer, so there won't be rounding
errors or other problems that are possible with either Single or Double. I
can't tell you if this is the same in VB2008.

Try a look at this site;
I think your best choice will be xnumbersdll.zip
Unless intended for use in Vista.
Read warnings, first on the site.
http://digilander.libero.it/foxes/SoftwareDownload.htm
If you find it useable, then give the credit to the
two times Bob.

/senn