Applicability of open-source web mapping libraries for building massive Web GIS clients

Abstract

The increasing capabilities of web browsers and the growing spread of JavaScript have an impact on the development of web-based GIS systems. While in traditional Web GIS applications the client-side component is only responsible for creating representation models, modern geographically enabled JavaScript libraries have extended capabilities, making them capable of doing extensive tasks, like complex geographical analyses. This paper identifies the most capable libraries for being the basis of a Web GIS client (Cesium, Leaflet, NASA Web World Wind, OpenLayers 2, and OpenLayers 3) and compares them. The libraries are compared by their GIS feature coverage and some quality metrics. OpenLayers 3 is identified for being the most capable library by supporting nearly 60% of the examined GIS features, its small size, and moderate learning curve. For comparing the learning curves of JavaScript libraries, a new metric named Approximate Learning Curve for JavaScript is proposed, which is based on other software metrics.

Appendices

Appendix 1: JavaScript routine for measuring the exposed functions of a library

Appendix 2: Detailed support table of the candidate libraries

Category	Cesium	Leaflet	NASA WWW$^\mathrm{a}$	OL 2$^\mathrm{b}$	OL 3$^\mathrm{b}$
Rendering
Hardware acceleration	1	0	1	0	0.5
Render geometry	1	1	1	1	1
Render raster	0	0	0	0	0
Render image	1	1	1	1	1
Blend layers	1	0	0	0	0.5
Formats—vector
ESRI shapefile	0.5	0.5	1	0.5	0.5
KML	1	0.5	0.5	1	1
GeoJSON	1	1	1	1	1
WFS	0	0.5	0	1	1
Write transaction	0	0.5	0	1	1
Formats—raster
GeoTiff	0.5	0.5	0.5	0.5	0.5
Arc/Info ASCII GRID	0	0	0	0	0
WCS	0	0	0	0	0
Formats—image
JPEG	1	1	1	1	1
PNG	1	1	1	1	1
WMS	1	1	1	1	1
Formats—image—tile service
WMTS	1	0.5	1	1	1
TMS	1	1	0	1	0.5
Slippy map	1	1	1	1	1
Google maps	0	0.5	0	1	0.5
ArcGIS REST API	1	0.5	0	1	1
Bing maps	1	0.5	1	1	1
Database—connection
PostGIS	0	0	0	0	0
SpatiaLite	0	0	0	0	0
MySQL	0	0	0	0	0
Database—functionality
Using DBMS	0	0	0	0	0
Query/filter	0	0.5	0	1	0
Query language	0	0	0	0	0
Data—pre-process
On-the-fly transformation	1	0	1	0	0.5
Read attribute data	1	1	0	1	1
Z, and M coordinates	1	0	0.5	0	1
Geometry types	0.5	0.5	0.5	0.5	0.5
Spatial indexing	0	0	0	0	1
Geometry validation	0	0	0	0	0
Geometry simplification	0.5	0.5	0	0	1
Attribute table	0	0	0	0	0
Data—conversion
Interpolate	0	0	0	0	0
Raster to vector	0	0	0	0	0
Vector to raster	0	0	0	0	0
Data—manipulation
Update attribute data	1	1	0	1	1
Update geometry	0	1	0	1	1
Field calculator	0	0	0	0	0
Add/remove layer	1	1	1	1	1
Change layer order	1	1	0.5	1	1
Typed layers	0	0	0	1	0
Data—analysis
Basic geoprocessing	0.5	0.5	0.5	0.5	0.5
Topological analysis	0.5	0.5	0.5	0.5	0.5
Modify image	0	0	0	0	0.5
Modify raster	0	0	0	0	0
Raster algebra	0	0	0	0	0
Classification	0	0	0	0	0
Convolution$^\mathrm{c}$	0	0	0	0	0
Write WPS request	0	0.5	0	1	0.5
Projection
Transform vector	1	1	1	1	1
Warp raster	1	0	1	0	1
Well-known projections$^\mathrm{d}$	0.5	0.5	0.5	0.5	0.5
Custom projections	0	0.5	0.5	1	1
Interaction
Draw features	0	0.5	0	1	1
Modify features	0	0.5	0	1	1
Snap points	0	0.5	0	1	1
Modify view$^\mathrm{e}$	1	0.5	1	0.5	1
Select features	1	0.5	0.5	1	1
Query	0	0.5	0.5	0.5	0.5
Measure	0	0.5	0	1	0
Change time	1	0.5	0	0.5	0
Mouse coordinates	0	0.5	1	1	1
Representation—styling
Style vector	1	1	1	1	1
Style raster	0	0	0	0	0.5
Thematic maps$^\mathrm{f}$	1	1	1	1	1

Representation—Carto. e.$^\mathrm{g}$
Scale bar	0	1	0	1	1
North arrow	0	0	1	0	0
Legend	0	0	0	0	0
Graticule	0	0.5	0	1	1
Text box	0	0	0	0	0
Overview map	0	0.5	0	1	0.5

1. $^\mathrm{a}$ Web World Wind
2. $^\mathrm{b}$ OpenLayers
3. $^\mathrm{c}$ Also known as moving window (GRASS) and focal statistics (ArcGIS)
4. $^\mathrm{d}$ Projections in the EPSG database
5. $^\mathrm{e}$ Ability to pan, zoom, and rotate the map
6. $^\mathrm{f}$ Ability to create choropleth and proportional symbol maps from vector data
7. $^\mathrm{g}$ Cartographic elements

Appendix 3: Example JavaScript function