GIS

OWS are OGC Web Services. I would like to find OWS on specific topics by entering a keyword. What OWS search engines are out there? With your help, I'm trying to compile a list (now edited witinks provided by ThomasG, thank you!)

• GeoPole by SourcePole AG
  
• OWS search engine by Benjamin Chartier
  
• Google
  
• GeoMeta
  
• link list at OSGeo wiki
  

The objective was to split a line into segments with another line. Here's what I came up with, the first argument being the source line, the second the line used for cutting the first

SELECT astext(geom) FROM dump ((
SELECT ST_AsText(ST_FORCE_COLLECTION(ST_Difference(
'LINESTRING (5 5, 3 3, 2 2, 0 0 )'::geometry,
'LINESTRING ( 2 0, 0 2 )'::geometry
)))
))

The result is a set of linestrings.

LINESTRING(5 5,3 3,2 2,1 1)
LINESTRING(1 1,0 0)

Absolutely trivial, but maybe you can use it.

Today I had some free time on my hands, and created a useless mash-up. The idea is to display the filming locations of a certain movie on a map. Here's what you have to do:

- copy this link into your bookmarks toolbar
- visit the IMDb and navigate to a movie, for example "Into the Wild"
- click the link in the bookmarks toolbar

You are then being redirected to a website that displays a map like this



The PHP parser gets the location names from IMDb, sends these names as query strings to the Yahoo! Geocoding API, which the returns coordinates the are being displayed by OpenLayers. It's needs some fine-tuning, but you get the idea

And finally, the credits: This mash-up uses Openlayers, the Yahoo! Geocoding API, jQuery, jQuery Progression, jQuery BlockUI and OpenStreetMap data.

As Mapbender makes heavy use of the jQuery library, I thought it would be nice to use its plugin system for adding control elements to Mapbender map applications.

Currently, all Mapbender application elements (like zoom, pan etc.) are configured with Mapbender's own administration interface, which is an extended HTML editor: Each element is represented as an HTML DOM element with additional information like styles or behaviour. For example, a zoom button is nothing but an img tag with a CSS style (e.g. position in the application) and behaviour (a JavaScript function that adds events like onclick to the button).

You can also customize an element by passing element variables to it. This is nothing but a set of key/value pairs that are entered into another administration form, and which will be interpreted by the JavaScript.

An easy way to standardize the interface between the Mapbender administration and Mapbender application elements would be to use jQuery's plugin system. The syntax of jQuery is quite easy:

$(selector).plugin(options)

This would translate to Mapbender like this

$(HTML DOM element).behaviour(customization)

The Mapbender element JavaScript would look like this

$.fn.element_id = function (options) {
    return this.each(function () {
        // add behaviour here
    });
}

I think if we adopted this style of creating application elements, developing for Mapbender would be much easier.

As far as I have understood, GeoExt is doing something similar, but is using Ext instead of jQuery. But I'm not sure if there already is a backend to configure GeoExt plugins.

A "slippy map" is an interactive map with an intuitive user interface: The user can simply navigate around in the map, while new parts of the map (typically, tiles) are added dynamically. The map as a whole doesn't have to be reloaded, which gives the user the illusion that the map is fully available all of the time.

Mapbender is only able to display untiled map services (OGC WMS). Due to this restriction, it is not feasible to preload or cache images efficiently: There are no rigid zoom levels, and this makes the amount of possible map images infinite. The downside is that moving around on the map is not a seamless experience for the user. There are a few things that can be done about this, unless you abandon untiled WMS and use tiled maps. I would recommend using OpenLayers for tiled maps.

But this is what can be done: The old map image could be kept until the new one is loaded, to avoid a blank map. The old image could also be adjusted while the new map is loaded, so the extent of the old map is correct in respect to the extent of the new map: Practically, this means that the old image could be moved when navigating via "center" or the "navFrame" buttons, or could be enlarged or shrinked when zooming.

Each new map requests generates a new div, which is positioned according to the old extent of the map. Once the map is loaded, the div becomes visible. The animation effects are simply done by jQuery's animate function. After the animation is finished, the div of the previous map is removed.

Here's a demo:


To include this demo on your website, use this bit of HTML.

Please tell me what you think about it. Maybe you have some good ideas how to improve it. Please file bugs under this ticket at the Mapbender bug tracker.

FOSS4G is the international conference for Free and Open Source Software, to be held in Sydney Australia, 20-23 October 2009. It attracts the best international developers and companies for Geospatial Open Standards and Open Source, as well as agencies deploying Spatial Data Infrastructure solutions, largely from the government sector.





More details about FOSS4G's comprehensive program can be found at: http://2009.foss4g.org
You can view this press release with images at: http://wiki.osgeo.org/wiki/FOSS4G_2009_Press_Release_10


Media Sponsors

- Position Magazine: http://www.positionmag.com.au/
- Geoconnexions Magazine: http://www.geoconnexion.com/
- Directions Magazine: http://directionsmag.com/

Today I looked a little more into OpenId and OAuth, and how they fit together.

I watched that lecture by Joseph Smarr from Google I/O, "OpenSocial, OpenID, and OAuth: Oh, My!", which was very interesting, but so full with new information that I will need to watch it again. A review will follow. But I remembered OAuth was used in a mash-up called Wikinear that I think is definitely worth talking about.

Wikinear is a mash-up which displays placemarks on a map. These placemarks correspond to locations of Wikipedia articles near a given position. Imagine you're a tourist, and walk through a city. By passing your location to Wikinear via GPS, you will discover interesting sites nearby (at least they are relevant enough to have their own Wikipedia entry). It looks something like this:



Wikinear is based on four components, the most important being a reverse geocoding web service by Geonames.org (findNearbyWikipedia). This returns Wikipedia articles within a distance around a given location (Geonames.org also offers to retrieve Wikipedia articles within a bounding box). Here's a sample request.

The map is just a static Google map. You don't need JavaScript when displaying a static map, this is just a static image specified by a few parameters. Obviously, you need a Google Maps key.

The third (and I think, most interesting) component is FireEagle, which acts as a location broker (Here's a presentation). Applications feed location data into FireEagle, and the user decides, which data is shared with which applications. You can, for example, let your GPS-capable mobile phone update your location data in FireEagle (you can also log in manually and enter an address, which will be geocoded; this is what I did in order to get the screenshots). Then you can register Wikinear with FireEagle, which will look something like this:



The communication between Wikinear and FireEagle is secured via OAuth (Here's a description). OAuth shares confidential information tokens while protecting account credentials. I think this means, that you can let the Wikinear application use your location data while logging in at FireEagle only. But I will surely rewatch the lecture mentioned at the beginning and share any new insights.

I tried to sum it all up in a diagram.