request for comments

[Henrik Brautaset Aronsen]

Den 22.08.12 12:43, skrev Lutz Vieweg:
> On 08/22/2012 07:38 AM, Dirk Hohndel wrote:
>> e.g., for a trip with four dives in Hoodsport:
>>
>> <dive number='4' tripid='-21' date='2010-12-04' time='10:38:00' 
>> tripentry='TRUE'>
>>    <location>Sund Rock, Hoodsport, WA, USA</location>
>> </dive>
>
> From a data model design perspective, it's a sin to abuse a <dive>
> entry to represent something completely different (a trip), to be
> distinguished just by some parameter.
>
> It may be pragmatic to do so for the moment, but once you add
> more types of entries to the file, you may end up with messed stuff like
>
> <dive number=... tripentry='FALSE' fooentry='FALSE' barentry='TRUE'>
>
> I'd opt for a separate <trip ...> entry.

I agree completely.    I don't like the "magic" negative dive numbers 
(from a different proposal), and neither the messy "tripentry" variable.

Den 21.08.12 17:01, skrev Jef Driesen:
> I think automatically organizing dives into trips is a bad idea. It 
> might be convenient if you are diving mainly on holidays, but it will 
> be really annoying if you are diving locally on a regular base. For 
> example, I do the majority of my dives at local dive sites, and I dive 
> there quite regularly. Thus, with the heuristics of starting a new 
> trip with an surface interval less than x days, I would end up with 
> many mini trips, while the logbook should contain no trips at all.

And I agree with Jef as well: I don't like the automatic trip generation 
at all.  It would create trips based on day-to-day local dives which 
wouldn't make sense.

But explicit trip creationg would be cool.

Cheers,
Henrik