oship-dev team mailing list archive

[Tim Cook <timothywayne.cook@xxxxxxxxx>]

On Fri, 2010-11-12 at 00:24 -0200, Diego Manhães Pinheiro wrote:
> First of all, I've seen litte progress on tasks assigned for two months.
> So, I ask with great humility: Eduardo and Wagner, do you have any
> detail that stuck you on code implementation ?

Thanks for staying on top of this Diego. 

> At the moment, I'm stucked on specific detail about a DvDate object: The
> ISO8601 interface defined at openEHR specifications define some rules to
> validate a string date. The format must follow the gregorian calendar,
> which means that the most old year supported is the 0001 year. In other
> hand, the method described on the openEHR specification that validates
> an year considers 0000 as a valid year, but the year 0000 doesn't exists
> in the gregorian calendar. The ISO 8601 specification supports years
> before christ death, but the gregorian calendar don't. So, how could I
> fix that ?

First and foremost be aware that ISO standards are in most cases
guidelines that are seldom fully implemented as written.  In particular
8601 is quite exhaustive in the coverage of datetime functionality.  A
more realistic approach was taken by W3C in creating an implementable
profile solution.  There is a pretty good reference to the various
profiles that was specifically written for developers.  It is not a
normative document at all but it does reference the various 8601
profiles: http://www.hackcraft.net/web/datetime/ The openEHR
specifications are in error considering 0000 as a reference point
(IMHO). I will post a query on the technical list.  But for now I lean
heavily towards the W3C profile of 8601.

> The simple solution that I found is don't validate days on
> months(because the gregorian calendar don't support it) if is the year 0.
> I will be glad on any other better solution. Please, don't throw up the
> possibility that I'm wrong on any detail described here. I'm not an ISO
> 8601 expert, so any idea will help. :)

I doubt there is an 8601 expert.  :-)  One more point is that openEHR
"added" to the DV_DURATION specification:

"Note that a deviation from ISO8601 is supported, allowing the ‘W’
designator to be mixed with other designators. See assumed types section
in the Support IM." 

The reasoning used for this makes no sense.  The issue is really one of
display convenience for obstetricians because they like to think in
Weeks. I lost this battle when on the openEHR ARB but it is not allowed
in 8601 so it should not be included as an option.

Unless of course someone on the list sees that it should be included;
and has a good reason for it.  IMHO, it should simply be calculated into
weeks if that is that they want to see on the screen. 

Thanks,

Tim

Attachment: signature.asc
Description: This is a digitally signed message part