🚀 go-pugleaf

In article <JENNINGS.92Dec17144132@antwrp.antwrp.gsfc.nasa.gov>, 
jennings@antwrp.antwrp.gsfc.nasa.gov (Don Jennings) writes...
>In article <9212171630.AA05949@tetra.gsfc.nasa.gov> William Pence writes:
>   My reaction to this is that it is not very useful to put information in a FITS file
>   that only a particular machine will be able to interpret.  One of the main purposes of 
>   FITS is to make the information in the file portable to any machine.  So I think
>   you have 2 choices.
> 
>The precision is significant, in that we must be able to repoduce the original
>data byte for byte. Breaking the VAX R*8 up into an IEEE R*8 + remainder would
>still introduce roundoff error.

I just want to cast a vote that stuffing part of the VAX precision into
the IEEE format (with an auxiliary byte) seems like questionable  style
to me.  It would be much better to split the number up, as someone
suggested, into an exponent and mantissa, then store those as integers.
Extended precision formats you make up yourself from scratch, cleanly,
should be easier to deal with than ``IEEE-plus.''

Perhaps it would be better for user-friendliness to make the exponent

       fix(log10(stuff))    instead of   fix(log2(stuff) + offset)

with the corresponding change in the mantissa.


Robert S. Hill                  
bhill@stars.gsfc.nasa.gov