[OpenLayers-Dev] Re: [OpenLayers 3] code sprint summary, and request for comments

[Eric Lemoine]

On Tue, Sep 21, 2010 at 7:54 PM, Peter Robins
<openlayers at peterrobins.co.uk> wrote:
> On 21 September 2010 12:34, Eric Lemoine <eric.lemoine at camptocamp.com> wrote:
>> Right. Do you think the proposed design doesn't address this scenario?
>
> not sure :-) Google's overlay layers are commercial layers but should
> not be part of the Google master group, as then you wouldn't be able
> to overlay them. As long as the proposal can handle that, no problem.

Commercial layers being in the master group is the default, but this
is configurable.

>
>>> The layerswitcher was just an example. Vectors would be another
>>> example of the advantage of separating OL layer from source. A vector
>>> layer might take data from several different sources, which might be
>>> in different projections - the projection is a property of the source
>>> not the layer. The vector layer doesn't need a projection property, as
>>> it will always be that of the map, as determined by the baselayer (or
>>> whatever you want to call it). The read() function should use the
>>> projection defined in the source file (or the default for formats like
>>> kml that are always in one (non-)projection), and do the appropriate
>>> transforms. The current requirement to specify internal/external
>>> projections shouldn't be necessary.
>>
>> Our BBOX, Fixed, and Save strategies already have support for this. If
>> the vector layer has a projection defined then the strategy will do
>> the transformation. Are you thinking of something else?
>
> yes. Setting projection on the layer doesn't help in the case where
> you have more than one source in different projections (though I agree
> that's probably not very common). IMO it also confuses things, as it's
> not the projection that the feature coordinates of the layer are in,
> which will always be the same as the map projection.IMO it would be
> better if this were renamed sourceProjection or some such. If you add
> the ability to reproject maps, then the projection of the vectors
> currently in the layer will have to be transformed from the old map
> projection to the new. The current layer projection property won't be
> of use for this.

I fully agree with this. If you care enough about it then feel free to
create a ticket with OpenLayers 3.0 for the milestone.


-- 
Eric Lemoine

Camptocamp France SAS
Savoie Technolac, BP 352
73377 Le Bourget du Lac, Cedex

Tel : 00 33 4 79 44 44 96
Mail : eric.lemoine at camptocamp.com
http://www.camptocamp.com