On 2016-08-05 14:17, Dave F wrote:
> On 05/08/2016 12:11, Colin Smale wrote:
>
>> On the other hand, consumers will have to deal with the MP model anyway, so
>> having multiple models for the same thing is actually adding to the
>> complexity.
>
> MP's were never conceived as a blanket 'cover all' solution.
> "I did it for a couple of schools, so I thought I'd do it for the rest of
> them" is not a valid tagging technique.
What I meant was, having established that some (many?) schools will need
to use the MP model, all consumers (for this data) will need to be ready
to process MPs anyway. A single polygon consisting of a single way is
just a degenerate case of an MP. Modelling all of them as an MP saves an
"if" statement and a second set of logic to process the data.
The single, closed way and the trivial MP containing such a way are
entirely semantically equivalent. There is no need to have two ways of
representing exactly the same information.
//colin
_______________________________________________
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb