[OpenLayers-Users] proj4js question

bartvde at osgis.nl bartvde at osgis.nl
Wed Jul 23 09:09:09 EDT 2008


Yeah, that's what I was thinking as well (and what I changed in my version
of proj4js), the transformed check is doing more bad than good IMHO, so I
got rid of it.

Hopefully Mike can comment on this.

Best regards,
Bart

> On Wed, Jul 23, 2008 at 02:55:43PM +0200, Patrick Valsecchi wrote:
>> On Wednesday 23 July 2008 14.46:22 bartvde at osgis.nl wrote:
>> > I am transforming a geometry from EPSG:28992 to EPSG:4326, and after
>> that
>> > to EPSG:23031. In the last step the geometry does not transform
>> anymore.
>> >
>> >         if (point.transformed) {
>> >           this.log("point already transformed");
>> >           return;
>> >         }
>> >
>> > Apparently proj4js thinks the point has already been transformed.
>> >
>> > Who is responsible for managing the transformed property? Since it
>> does
>> > not seem possible to reproject geometries more than once currently?
>>
>> Hi,
>>
>> Just do that before the second transformation:
>>   delete point.transformed
>>
>> And you'll be good to go.
>>
>> I don't know why proj4j changes the passed object instead of creating a
>> new
>> one. It's a strange design decision, but it's like that...
>
> It's not strange at all.
>
> I have a 10,000 point vector I've downloaded from a server. I
> constructed it, I'm ready to use it. First I need to transform it.
>
> Creating 10,000 objects (OpenLayers.Point objects) is expensive --
> specifically, object creation in a tight loop like that is probably
> about 40% of the total cost.
>
> Scale it up another order of magnitude, and the difference moves from
> hundreds of milliseconds to seconds.
>
> proj4js is doing the right hting by not creating a new object.
>
> The only thing I don't understand is why it doesn't want to reproject
> something ith as already projected. I think *that* is a weird decision,
> personally: But I expect the reason is "Let's not help the user shoot
> themselves in the foot."
>
> My guess would be that the right thing to do is to get rid of this
> 'already projected' check, and expect that Applications will do the
> right thing.
>
> Regards,
> --
> Christopher Schmidt
> MetaCarta
> _______________________________________________
> Users mailing list
> Users at openlayers.org
> http://openlayers.org/mailman/listinfo/users
>





More information about the Users mailing list