Tuesday, July 28, 2009

Re: [Geopriv] geo: URI: what about uncertainty?

>>>>> Alexander Mayrhofer <alexander.mayrhofer@nic.at> writes:

>> The instant you raise the point of significant figures in this
>> forum, you are asking for a world of pain. I'll give you a running
>> start.

>> I'd actually like to see text that says that 'geo:1,2' ==
>> 'geo:1.0000,2.00000' to avoid this issue/trap entirely.

Well, it will allow for better scanf (), printf ()
compatibility, at the least.

> Actually, i have been personally approached by an AD yesterday,
> requesting that i add exactly such text to the draft. I will do that
> in the next version to get that document forward.

Could you please consider a clarification like the following as
well?

[...] In particular, the number of trailing zeros MUST NOT be
treated as an indication of the precision of the values given.

Perhaps a generalization should then follow immediately? Like:

The digits forming the values MUST NOT be unconditionally assumed
as all being significant.

--
FSF associate member #7257
_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv