Next round of api improvements.

[Jef Driesen]

On 24-11-12 13:26, Artur Wroblewski wrote:
> On Wed, Nov 21, 2012 at 4:57 PM, Jef Driesen <jefdriesen at telenet.be> wrote:
> [...]
>> Let's say we define the meaning of the temperature as the minimum water
>> temperature, which is probably the most common value. But then you
>> happen to have some dive computer which only provides the average water
>> temperature. Since that's all we have, that's what we give to the
>> application and goes into your logbook. Is the small difference in
>> semantics really that important, considering we don't have any other
>> info?
>
> Another example is Reefnet Sensus devices and their time implementation
> (IMHO, quite reasonable solution as it is... hm... headless device? no buttons,
> no display). It reports time according to its internal clock, so to
> calculate the
> real time you need the start of the internal clock, time of dive according to
> the internal clock and real time of data download.
>
> I believe libdc will handle above with no problem while parsing data during
> download from the dive logger.
>
> But when you want to reparse the data (which, I hope, will always be possible
> with libdc), then an application needs to be aware of Sensus' time semantics.

Re-parsing (or offline parsing as I usually call it) will remain supported. Of 
course the application will have to store the device and host clock timestamps 
along with the data. But that's why we have the DC_EVENT_CLOCK event, and that 
won't change. In the current design, you pass the two clock timestamps to the 
reefnet_sensusultra_parser_create() function. In the new design that will be 
done in a similar way, except that for the immediate parsing this will all be 
taken care of internally.

Jef