Stefan Knecht, co-founder and CIO of United Maps, got in touch with me providing information about his company and their products. United Maps works hard on adding more value to existing maps, as we know them on Google Maps or in automotive navigation systems, and create digital maps for humans. Considering the increased popularity of GPS enabled mobile devices and the given potential of pedestrian navigation systems, it seems to be the right thing to do these days.

Over at Vector One and at United Maps‘ website (blog) you’ll find detailed information about their product and vision.

I took the chance and asked Stefan some more questions, see below.

Q: United Maps creates digital maps optimized for pedestrian use, a perfect addition to many mobile mapping applications. Who would United Maps consider as primary target group? Is your focus rather on the white labeled map as data product or are you working on API like services to attract individual developers for instance too?

Stefan: I’d like to reframe “target group” to something more universal like “use group”. At the time being, we concentrate on delivering what we carry in our company name: a unification of maps, attributes and use cases to enable mobile people finding their way and discovering things around them.

So the focus is on comprehensive, nationwide and B2B data products rather than on public APIs and just another mash-up. We’re not mashing-up what’s already out there — we try to drill deeper and possibly beyond what’s easily visible on the web.

Q: The OpenStreetMap foundation is currently working on a new licensing model: ODbL should basically allow OpenStreetMap features and copyrighted map features being held in the same database. Have you considered OSM-integration in United Maps?

Stefan: First of all: OSM does a great job, all kudos to them. The recently completed dataset of Hamburg is incredibly good. I wonder how OSM will perform in “flat world”, outside of larger cities and how OSM will be able to scale into less populated and geek-prone areas.

To answer the question and as far as I can judge from the ongoing debates within the OSM community: the modularity of a dozen CC license types shouldn’t be brought into ODbl. The legal situation already is far too complex - and it doesn’t become easier with just another set of derivative licenses and constraints to consider.

Q: I believe gathering detailed cadastral maps across Europe can easily turn into an exhausting process - different legislations, different mapping traditions and INSPIRE implementation has just started. Do your GIS experts consider other and maybe easier accessible sources, such as vectors derived from commercial EO data, rather than official public data to “fill the gaps” in Europe and push United Maps rollout forward?

Stefan: One of our goals is to match INSPIRE specifications on a base level to enable users of our United Maps gather and aggregate data on top. For other data sources besides federal information: any valuable source that can deliver coverage for a given set of national boundaries is welcome and might be licensed and matched with the data we already have. We’re positively testing options - and expectedly, both data availability and legal constraints change at every administrative border … or any 150 miles in Europe.

Q: Nokia Maps is probably a serious competitor for United Maps. As far as I know Nokia Maps, their approach is to provide landmarks instead of precise building footprints to support orientation or even suggest shortcuts through buildings for pedestrian navigation. Where would you see the main difference to Nokia Maps or what aspect do you think makes United Maps the better choice for pedestrians?

Stefan: It would be impudent to name United Maps as a competitor to Nokia Maps. Nokia Maps is a B2C product and naturally powered by Navteq data. For the time being, United Maps is in a B2B space.

I don’t see that precision of footprints is a real issue: it’s rather the availability and rollout of supplemental data to enhance the usage experience on Nokia Maps. If landmarks are helpful - why not integrate them? I don’t see us producing 3D-mockups for a simple reason: if you’re a human on the move, trying to orientate yourself on the 3-inch-screen isn’t really simplified by 3D-models that you rarely see in entirety in the urban jungle. If 3D-models remain picturesque building hulls they act as visual landmarks. The pedestrian shortcuts through buildings can only be produced with a topologically closed and hence routable network beyond — and this ultimately is, what United Maps does: gather content, attributes and pathways that are relevant for people outside of cars.

Q: You’re partnering with the Technical University of Munich. How important is the scientific input for United Maps? Is United Maps a research project?

Stefan: United Maps draws from the research we commissioned at TUM before we started the company. We repurpose the initial scientific results into a commercial setting and take academic aproaches onto a industrial scale. The scientific input is most valuable and will be perpetuated to specific domains and settings. We’re just developing a multimodal pedestrian routing application that seemlessly routes you back and forth through automotive traffic and mass transit alternatives.

The mother of all Q: Will there be an iPhone version of United Maps?

Stefan: United Maps does better, hyperlocal maps at large. The iPhone and all other smartphone devices will use the web for mapping and possibly web services navigation. If there’s a business case for a tailored iPhone application, we’ll do that in-house or offshore it to a partner. We’ll have Germany ready as comprehensive hyperlocal dataset in April, then Austria and Switzerland — everything beyond is subject to change and upcoming partnerships. And we’re naturally open to partnerships of any extent.

United Maps presentation at the Telematics Munich Show

…is something you may consider turning off if you’re unhappy with VMware Fusion’s performance.

As for ArcGIS Desktop, it works noticeable faster in VMware Fusion if the entire workspace is moved inside the VMware image instead of accessed via “Shared Folders”. ArcMap feels snappier and geoprocessing runs about 35% faster as my quick benchmark showed. Performance issues caused by “Shared Folders” is mentioned at an ESRI Discussion thread too.

My benchmark test was just a geoprocessing task I needed to do for a project, executed in 3 different workspace environments:

#1 - inside the VMware image
#2 - on an external USB harddisk, mounted in Windows
#3 - on an external USB harddisk, mounted in Mac OS X and accessed through VMware Fusion’s Shared Folders feature

The task was to intersect 2 layers, everything done inside a File Geodatabase:

Layer A: 11,932 features (20,4361 vertices)
Layer B: 3,100 features (arcs from point buffering)

The intersect-process returned a 300MB feature class containing 952,265 features and 5,724,810 vertices. Below is the chart showing the time needed to complete the task for each workspace environment.

ArcGIS geoprocessing task performance in different workspace environments

In figures, option #1 took 8 minutes 57 seconds, option #2 9 minutes 3 seconds and option #3 needed 13 minutes 29 seconds to finish.

Clearly, the performance bottleneck in VMware Fusion is “Shared folders”. No doubt, it’s a handy feature and makes file sharing between host and guest-OS very easy, but for performance reasons you better turn it off.

Once in benchmarking-mood I ran the same task with increased memory and 2 CPUs. Out of curiosity, just to see the effect of more memory and CPU power. My standard setting for VMware Fusion is 1 CPU and 512MB RAM allocated to the guest-OS, which turned out to be the best setting for working at decent speeds in both host and guest-OS so far. The result for the same geoprocessing task with 2 CPUs and 1024MB RAM was 8 minutes 17 seconds. Little faster, but, because of experiencing a sluggish host Mac OS X, not worth it.

Despite performance I ran into another problem with “Shared Folders” and File Geodatabases a while ago. Well, I actually never verified that this problem is related to “Shared Folders”: a File Geodatabase corrupted while executing “compact database” in a “Shared Folder” workspace. According to that thread at ESRI it happened on network drives too. However, I experienced it only in “Shared Folders”, not in other workspace environments. Quite annoying bug though.

Simple, yet interesting: search Twitter for a place name, bundle returned tweets in a new stream and watch what’s going on there.

Compared to a pure geographic search, the semantic search returns tweets about that place, instead of tweets located nearby. The content of geotagged tweets isn’t necessarily directly related to the place, as the geographic search results show. A combination of both methods might be useful though.

What if you search for directions from place A to place B while you are only able to use a keyboard as input and navigation device on websites?

MAIN_web, an Austrian initiative for media accessibility, asked to try and navigate the web only by using the tab key. I tried to get directions on three popular mapping sites: Google, Microsoft Live Maps and Mapquest. That’s how it went…

Google Maps

Google Maps initially focuses on the search bar and you can start typing the search query right away. Use “to” between both addresses and Google reads it as search for directions between 2 places. No tab yet needed (”enter” triggered the search) to get the first results. Unfortunately there are more options for one of both places and Google suggested some alternative addresses. Hoping down to the right suggestion took about 33 tabs. After that, the right directions showed up on the map, together with textual descriptions beside. The problem then was that it’s not possible to look up detailed views of the route only by navigating with the keyboard though.

Microsoft Live Maps

The site doesn’t focus on the search bar. It takes 9 tabs to reach the point where you can start typing the search query. The search engine didn’t understand “to” as search for directions, so I had to enable direction search first: 7 more tabs. From there you hop 19 tabs around the page to access the “start” field and one more to enter the “end” address. One address wasn’t found immediately, but the suggested correct address was only 4 tabs away. As before, the result page shows the route on the map and descriptions, but it’s not possible to access further or more detailed information on certain route items by only using the keyboard.

Mapquest

Mapquest puts the focus on the search bar. For directions you have to use the form below on the page: 5 tabs away before you can start typing start and end locations in different form fields. As before, one address wasn’t unique and Mapquest offers several suggestions. Unfortunately there’s no way to access those by using the keyboard only, the page sends you around in circles in the header area. So no directions from Mapquest at all.

It’s an interesting experiment and gives a feeling about web accessibility. I checked some other (own) mapping sites too and I guess there are some things we should look into. At least the main information or purpose of a site should be accessible that way.

This is going to be a list-post, a must-have item on blogs as I read recently.

Last week I attended Richard Weait’s OpenStreetMap presentation at the Boston LUG. He gave a good overview on what OSM is about, why it’s so fascinating and showed how to participate. Even though I’m a bit familiar with OSM for some time now, I learned a couple of new things about OpenStreetMap:

1. North America is only covered superficial. It looks good on a small scale, large areas appear mapped in OSM, but as you zoom in and look at it more closely you’ll find many problems. I’m not sure if motivation would be higher if people start mapping from scratch and “create” their own maps of their places instead of working on existing data and fix errors. On the other hand, the AND data donation of Indian and Chinese road data (mostly of poor quality) helped to create awareness and interest in mapping those countries, as some people mentioned during the discussion.
2. Germany has the largest OSM community. Not surprising that this country is among the first ones where OSM is integrated into business models (e.g. Geofabrik, WhereGroup or GDV).
3. Open Database License: the idea of Creative Commons transferred to suit database licensing.
4. “Edit wars” taking place in Cyprus about the default place names, shell scripts vs. manual edits.
5. OpenPisteMap: an OSM side project for ski slopes which adds hill shading and contour lines to OSM. One might think that it’s Austria’s #1 mapping project. Well, I’m afraid that’s not the case. Apparently not many fellow citizens have ever heard of it because we look like a desert in terms of available ski slopes at OpenPisteMap. Shame on us!
6. If you discover OSM flaws and can’t fix them directly on the map, then just go to and add a note in OpenStreetBugs.
7. Another interesting OpenStreetMap routing service I didn’t know before. It even considers elevation profiles for cyclists or pedestrians where available.

The parts of Cambridge I know so far are mapped fine (was a MassGIS data donation). The only thing I’m doing here in OSM is adding bike lanes to existing streets in my neighborhood. Since I really enjoy exploring this place by bike, that’s probably the most useful map layer to me.

Another option to integrate rendered OpenStreetMap images in a GIS desktop client or web application is to add OpenStreetMap as OGC compliant Web Map Service. For the European continent such a service is now provided by the German WhereGroup.

The service is available as a free basic WMS (GetCapabilities), updated biannually, and two commercial versions, updated daily, for specific map customization needs and high performances.

OpenStreetMap is on the way to become a serious alternative map source in professional GIS environments. The question how the professional GIS community responds to Creative Commons licensing will surely open some interesting debates anytime soon.

My favorite ArcMap extension Arc2Earth got updated and brings now rendered OpenStreetMap layers into ArcMap:

Previous versions supported map tile layers import from Virtual Earth, Yahoo! or Ask. Certainly a great feature which allows easy access to good and up-to-date base maps in many regions. The only problem is that without proper licensing it’s not possible to use them for commercial purposes. I’m not a lawyer, but as far as I understand the term commercial purpose, only loading and viewing those layers in ArcMap in a business environment (e.g. at the office) can already result in a license violation. Good license deals might not be such a problem at the enterprise level, for small businesses who need those maps here and there, it is quite often an issue.

Being able to load OpenStreetMap instead of Virtual Earth, etc. into ArcMap removes a lot of those licensing headaches.

OpenStreetMap data can be used freely under the terms of the Creative Commons Attribution-ShareAlike 2.0 license.

As long as attribution and share alike work for you, you can basically do whatever you want with OpenStreetMap in ArcMap - copy, distribute, print, derive, etc.

There is a long list of other new features and fixes that come with the latest release of Arc2Earth. OpenLayers as additional viewer for exported map tiles is one of those which caught my attention.

The map development team at Flickr released some interesting new additions to their geo API: shapes - not yet real ESRI Shapefiles, even though they’re on their way (see code.flickr for more information on that).

Flickr shapes are, as I understand it, polygons of aggregated point clouds of photo locations sharing the same location name. For geotagging, Flickr uses a very smart method: once the user has placed a photo on the map, the system sets a place name next to the photo. At the same time Flickr offers name modifications if the user isn’t happy with the proposed name. A list of name alternatives shows up where the user can pick the one which sounds best. That way Flickr constantly receives user feedback on its geodata and can continuously refine its geoname system.

On the other end Flickr makes the collected data through its API available again (see flickr.places.getInfo). I queried Vienna and some neighborhoods to see what the shapes look like in the city I know best. Surprisingly the city boundary is more accurate than I’d have thought. Keep in mind that it’s just the result of people geotagging photos and not surveying an administrative border.

The red line is the Flickr shape, the white line is the city boundary in Google Earth.

To obtain proper Flickr shapes on smaller neighborhoods, a certain critical mass of geotags needs to be achieved. Especially tourist hotpots turn out to be a potential pitfall: there is a high share of users geotagging without good local knowledge. Locals, who usually know the area better, won’t move around tourist attractions and take pictures that much. The relatively small amount of more accurate geotags done by locals will vanish in the mass of inaccurate geotags.

The yellow shape is Stephansdom, probably mostly tagged by thousands of tourists. Although Stephansdom is supposed to be the city center, it’s still only a square around a church within the neighborhood Innere Stadt, the green shape. As the picture shows, the relation and location of both shapes is slightly shifted.

I think some sort of ranking mechanisms can help here - a proper method to determine how accurate and trustworthy a name and corresponding geotag are.

Anyways, the idea of crowdsourced geonames on Flickr is interesting and it’s generally fascinating to watch the development a photo gallery has gone through over the last years. [via geobloggers]

Second Light is a new development of MS research on Microsoft’s multi-touch device Surface. It basically allows to display additional information in some kind of hidden light layer above the Surface screen. That way, extra map layers, like labels, can be shown on top of a base map without covering geographic objects for instance. It seems to be an interesting option for visualizing and exploring geographic content - makes me wonder when we see Virtual Earth demoing on Surface.

The video below gives a quick demo of how Second Light works.

[via rolf generated content]

Update:

The All Points Blog covers Second Light too, along with other new GeoGoodies from Microsoft.

That’s what I actually had in mind: creating a local Sandbox in VMware for testing GeoServer in conjunction with other FOSS GIS software. As it turned out, it’s not necessary to set such a system up by myself. Ricardo Pinho did it already and made a VMware image already available:

GISVM is intended to be a full-feature GIS Workstation based exclusively on free GIS software: PostgreSQL, PostGIS, GeoServer, Mapserver, FWTools, QGIS/GRASS, gvSIG, uDIG and Kosmo, on Ubuntu Desktop.

Great work! Exactly what I wanted. It’s a huge time saver. GISVM gives you a feel what FOSS GIS software can do for you, without going through any installation and configuration hassles.

GISVM works well as sandbox in VMware fusion here on the MacBook (1 processor and 512MB RAM enabled) , although it asked for upgrading the image file and installing VMware Tools for better performance.