VOunits draft

[Rob Seaman]

On May 26, 2009, at 12:56 PM, Anita M. S. Richards wrote:

>> So will folks have to convert 1 megaparsec to 30.9 Zettameters (Zm)?
>
> One folk will write a piece of code (or has already done so) which  
> everyone else can use seamlessly without even realising.  If it does  
> not give the game away by converting 30.9 to 30.8888888756789...

So the answer to my question is yes?  The most fundamental of the  
numerical user issues are errors introduced by back-and-forth unit  
conversions.  1.000000000 megaparsecs turns into 30.85.. Zm, which are  
truncated back to 0.999832... megaparsecs at the other end.

VO users will expect that an input table with a list of distances in  
integral numbers of parsecs will not be turned into a bunch of  
floating point numbers suffering from various truncation and rounding  
errors.

> er, don't we all want everyone to use standards which the IVOA has  
> adopted?

We want them to use IVOA standards because the requirements of their  
own projects make specific IVOA standards the most appropriate for  
their purposes.  This appears to be the case with VOEvent, for instance.

> And NB the Units workpackage is only talking about providing tools  
> for conversions which are the same for any given data - e.g. not  
> most magnitude systems...

If I understand your point, this would eliminate all the STC use  
cases, too.

Rob