Re: [xml-dev] ANN: a portable data component -- length

[Amelia A Lewis <amyzing@talsever.com>]

On Sat, 09 Apr 2011 12:00:20 -0400, Liam R E Quin wrote:
> On Sat, 2011-04-09 at 11:10 -0400, Costello, Roger L. wrote:
>> Here is the altitude of an aircraft, expressed in both feet and meters:
>> 
>>      <altitude>
>>          <feet>12000</feet>
>>          <meters>3657.6</meters>
>>      </altitude>
>> 
>> It is important that the two length values are consistent:
> 
> The right way to mark this up in most cases is to store only one value
> and convert as needed.
> 
> See Normal Form in database theory: information should never be
> duplicated. That way it can't get out of sync.

Agreed. My initial thought, on seeing the base post, was that this was 
"How to generate problems in data synchronization", not "portable data 
structures".

By including two values in different units, rather than a single value 
and relying upon the *well-known* formulae for conversion, all that is 
achieved is to create doubt as to the accuracy of any reported value.

As for the claim that the assertions will ensure that the data never 
gets out of sync, poppycock. The initial statement is that the 
information is 'portable', and yet this allegedly portable information 
relies upon some external code that enforces assertion validity? And 
what good do assertions do when these values *are* out of sync? Once 
they've gotten out of sync (that is, at any stage of the process at 
which someone fails to verify the assertions, or even any stage at 
which someone uses a different degree of confidence, or uses a 
different chain of conversions to generate values), *you don't know 
which one is correct*.

The following is valid:

     <altitude>
         <feet>12000</feet>
         <meters>4000</meters>
     </altitude>

... within a certain degree of confidence. 1 m equals one yard 
(rounding up; in some applications this is a perfectly valid 
conversion). Now, it's highly unlikely that such an application 
envisions measurements in thousands (at which point the 
less-than-ten-percent conversion error has made for very large 
differences), but ... in the above example, *which value is correct*?

Or, to quote the storied wisdom of the ancients: a man with one clock 
always knows what time it is. A man with two is never sure.

Amy!
-- 
Amelia A. Lewis                    amyzing {at} talsever.com
And now someone's on the telephone, desperate in his pain; 
someone's on the bathroom floor, doing her cocaine; 
someone's got his finger on the button in some room--
no one can convince me we aren't gluttons for our doom.
                -- Indigo Girls