Re: [Gnumed-devel] Measurements - adding and abbreviations

[Jim Busser]

On 2011-07-31, at 1:05 PM, Karsten Hilbert wrote:

> Please suggest a saner source of the time when no time is
> provided.

I suppose there is in postgres no field that

        requires date to be provided (but may optionally allow null)

        accepts a time portion (if inputted) but able to optionally allow a 
null time portion

??

> Use now().time ?

when the date *is* today, now().time might be reasonable (inputting a blood 
pressure measurement)

when the date is anything other than today, now().time is clearly irrelevant in 
which case I would…

> use 11:11:11.1111 to make it fairly apparent this is an
> assumed time ?

*******************

> Now, when the timestamp originated from parsers other than
> mx.DateTime it would with certain inputs know the actual
> input accuracy and thereby set the display accuracy to that.
> Hence it could appear as if only a date was used (w/o time
> part) - in which case most often the time part of now() was
> used. For the sake of consistence GNUmed will now force the
> display accuracy to "minutes" regardless of the parser that
> produced the timestamp.

it is a shame we have to include a time portion (assuming or arbitrarily adding 
now().time) in cases where now().time is orthogonal and in fact misleading.

??

-- Jim