This actually raises a problem when using text archives with floating point data - if you want to be able to round trip the values, then how many digits precision should you set the stream to? Particularly if you're saving a complex structure containing different floating point types of differing precisions?
Let me rephrase this somewhat: if I'm the author of a complex data structure that contains floating point data and I make it serializable, I have no way to ensure serialization proceeds correctly: instead that burden falls on the person creating the archive. For sure in small projects these will be the same person, but clearly not when "library writing". And yes, I realize that many compilers don't round trip floating point values even when you print "enough" digits, but at least they get close. The current default situation is to print just 5 digits, and that certainly doesn't get close! But, rant over for now ;-) Cheers, John.