[QUANTITY] doc consistency

[Jonathan McDowell]

Pierre,

: >  Many things are optional in the sense that if absent in the
: >  serialization, their value is a special value indicating  the absence of
: >  valid data. 
: Then unitless which correspond to the absence of data, because it is not needed,
: not due to uncomplete knowledge of the quantity object, does not match
: this def of default value.

Perhaps I used the word 'valid' wrongly... I didn't really mean 'valid',
I meant perhaps 'of the standard form'. 

Anyway the question is, if the special values of Unit are <unitless> and <unknown-unit>,
which should be the default?

I was arguing in favor of <unitless>, but probably it should be
<unknown-unit>, since then lazy providers will not generate large
amounts of data that is interpreted as <unitless> when really it is just
<unknown-unit>. 

So we could make the default 'unknown' and keep <unitless> as a special
value but not the default value.

The counter-argument is that <unitless> is really common, especially
for integer and string data. What do other people prefer?


 Jonathan