[OpenLayers-Users] Is processing GeoJSON slower than GML and KML?

[Tondo]

Hi lists,

I was testing how fast GeoJSON is and found out strange results...

I made 2 samples (simple feature) and downloaded three actual data samples
(state, region and county boundaries from DIVAGIS). All samples were
"reconstructed" by the OL.Format.Write() - so I made sure I am using "OL
compatible" data.

As expected, GeoJSON, WKT and Encoded Polyline were pretty fast while
working with geometry and feature with attributes:
http://jsperf.com/openlayers-format-reading-speed/2
http://jsperf.com/openlayers-format-reading-speed-attributes/2

After that I wrote tests for actual data (three layers, [1, 8, 79 features],
[70, 16, 18 attributes per feature class], [2520, ~1800, ~900 verticies per
feature]). You wont be able to run those since I don't have currently access
to any storage where data could be placed :( and I am testing locally.
However, test cases are the same:
http://jsperf.com/openlayers-format-reading-speed-large-feature-collectio/5

According to results GeoJSON is the slowest.

Are those test accurate? Am I doing something wrong, missing something
obvious?

Regards,
Tom



--
View this message in context: http://osgeo-org.1560.x6.nabble.com/Is-processing-GeoJSON-slower-than-GML-and-KML-tp5073040.html
Sent from the OpenLayers Users mailing list archive at Nabble.com.