february 2004 archives

Wednesday, February 25, 2004

Web mapping and usability

Once upon a time when your school atlas was too heavy to carry it to school for your geography class and road maps were too difficult to fold and road atlases never included the town plan of the settlement you just got lost in, indeed, when maps were still printed on paper, cartographers always made sure to include as many supporting tools as possible to navigate their products. At the back of an atlas there was an alphabetical index of setttlements or streets, the inside of the covers sported overview maps informing you on which page you could find which map. Dazed by all the colours and symbols? A legend or map key was there to help you out.

Granted, searching for a map of a specific location has improved a lot since then. No going back and forth between index, overview map, and map sheets to check the map reference of a location: just type in a town name or postcode and the online mapping website displays the correct map with an icon indicating the location of the town or postcode you just entered. However, there's more to usability than just that! Among cartography-related websites there's hardly any interest for this. Fortunately, on usability-related websites there's quite some talk related to search form usability for example, or how best to design a simple location map.

Now, back to the issue of search form usability regarding geographic queries. There are a lot of problems to overcome when we have a look at the various address details:

  1. Most mapping websites cover many countries, so you will have to select a country first to narrow your search. A typical approach is to have the home country of the mapping website pre-selected.
  2. The address structure, the order of address details differs from country to country, e.g. UK addresses put the town before the postcode, whereas Dutch addresses put it the other way around. Then again, postcode areas in Germany or France are so large that they're not very relevant for geographic searching.
  3. Geographic names, toponymes, are sometimes different across languages. Londen and Londres are not too different from London. But would you realise that Mons and Bergen is the same Belgian town? It doesn't stop with country names or town names either: even streets can have multiple names.

These idiosyncrasies aside, there's a lot one can do to improve the user experience. That usability is a serious issue becomes clear from an case study by Design Not Found. A locator usability report from the Nielsen Norman Group gives some recommendations with particular reference to store locators. Indeed, a recent publication on Dutch store locators clearly highlights the need for improvement, but geographic search forms are not limited to store locators. Widgetopia got some interesting comments on how to improve geographic searching on the FedEx Transit Time Calculator. Really, this is just the tip of the iceberg. There are really so much more usability issues that cartographers have to consider. Be sure to stop by again if this piece has made you curious!  permanent link for this entry

Monday, February 23, 2004

Free linking services for UK maps

Imagine you just bought a new mobile phone. You know, that geeky one you couldn't quite afford that comes with a camera? Right! Now, you have been taking pics all weekend and you are thinking of a geeky way to put all of them up on your website. Hey, you think: wouldn't it be cool to show the locations where I took these pics on a map? You read about other people doing similar pet projects on some mailing list. Then you suddenly wake up from your dream and realise: “I am living in the UK. Ordnance Survey. Crown Copyright”. Your dream has turned into a nightmare!

Not quite so. Both UK online mapping providers, Multimap and Streetmap, have a free linking service you can use to do just that: plot locations in the UK on a map! Both web sites display the URL to link to the current map view. This can be useful if you know the address of the location you'd like to map, but you don't know its longitude and latitude. Usually the quickest way is to enter a UK postcode in the search box:

Both URLs link directly to the public web sites of either of the online mapping providers displaying the map of SW11 6NU. Note that the coordinates in the URL provided by Multimap are less accurate! Below the map, Multimap also displays the OS grid coordinates, the WGS 84 coordinates in longitude and latitude, both in decimal notation and in degrees, minutes, and seconds. An OS Landranger map sheet reference completes the information. Streetmap provides a link to the Grid Conversion Results to obtain similar information. The Grid Conversion Results web page can be used to convert between various coordinate systems. How to operate this tool is described in the Conversion Tips.

If you require further customisation, Multimap provides a Automatic Link Generator. Supplying a UK postcode, or an OS grid reference, you can create a hyperlink for a web page with a custom title and scale. You can also opt to be listed in the Local Info section of the Multimap web site. Further information on customising hyperlinks and how to create links for areas outside the UK using WGS 84 coordinates in longitude and latitude is provided as well. Free linking is also covered in the Multimap FAQs.

You'd rather have a hyperlink to Streetmap? Check out their online link creation tool. It explains how to link by UK postcode and by OS grid reference. There's even a snippet of code to integrate Streetmap with Outlook: you can go straight from a contact to a map of where they live. The Streetmap allows for clever customisation:

  1. If a title is supplied in the, an optionally back title and back url can be supplied. The back title is a description of a hyperlink back to the original page and the back url is the URL of the hyperlink.
  2. As not to display any other local web sites on the map, you can supply a no local sites parameter.
  3. To customise the basic style of the web page, the body flags parameter can be supplied with typical HTML attributes for the body tag.

Unfortunately, the hyperlinks that are included to demonstrate the capabilities are all redirected: not quite the experience you'd like your visitors to go through. Also the hyperlink to go back to the original page is not clearly visible on the Streetmap page.  permanent link for this entry

Sunday, February 15, 2004

Meta comments

Having discussed online mapping in general since 2001, webmapper.net has recently focused on locative media and collaborative cartography quite a lot: democratisation of cartography. Projects have been highlighted and put in the context of general developments within the GI industry. Well, it's still all theoretical. Fortunately, there are people doing the real thing, such as Urban Tapestries. They have been through their first public trial in the Bloomsbury area in London. Based on their experience, they have identified five themes that are open for public discussion:

  1. Collaborative Cartography and Location Sensing
  2. Citizenship and the Public Commons
  3. Mobile and Pervasive; Spatial and Temporal
  4. Sensory Stimulation
  5. Filtering out the Noise

This discussion is surely worth having a look at. It is a great opportunity to comment on a revolutionary, location-aware commenting system. By the way, if you are curious what the lot at ETCON have been up to, you'd better check out the stuff at the locative ETCON wiki. UPDATE: Jo Walsh just sent out a summary of what went on in San Diego to the geowanking mailing list!  permanent link for this entry

Tuesday, February 10, 2004

A Geocities of GIS

More and more, people are thinking of ways to deal with location on the web. Of course, it has been mentioned before webmapping has become fairly commonplace. Furthermore, there are OGC specifications now that provide mechanisms to achieve interoperability between webmapping applications. Jo Walsh recently posted an interesting comment to the geowanking mailing list: if Geoserver (an open source implementation of the WFS specification) is the Apache of GIS, then what is the Geocities of GIS?

Together with Schuyler Erle, she has written an RDF aggregator for locative packets, enabling collaborative mapping applications to communicate. It's the foundation for the Collaborative Mapping Workshop at ETCON in San Diego that started yesterday. For more location-related events at ETCON, check out the links in this previous post.  permanent link for this entry

Thursday, February 05, 2004

Collaborative cartography is taking root

Some time ago, I was approached for a short interview in contribution to an article in Release 1.0 on the value of geography for business applications. But the article heralded an ever bigger development to take place:

Thanks to the constellation of technology that enables digital networks to be laid over the places of the earth, wherever we are we will be able to hear the human conversation that has occurred about that place - the history that occurred there, the aesthetics to be savored, the commerce transpiring at that very moment, recommendations offered by strangers and friends. The mute places of the earth are being given voice, and the voices are, of course, ours. Meanwhile, the places themselves are becoming digitally alive and are noting our presence, too.

David Weinberger is not the only prominent person to realise this development. In light of the upcoming Emerging Technology Conference 2004, Tim O'Reilly recently posted three inspiring observations regarding online mapping services to the geowankers mailing list:

  1. All of the “killer apps” of the first generation web, e.g. Google, Amazon, and eBay have started down the path of turning themselves into platforms, rather than just applications, except for the online mapping provider MapQuest.
  2. Of all the internet killer apps, MapQuest is also the only one that hasn't become dominant.
  3. Except for MapQuest, the internet killer apps are all having their users build the database as a side effect of their individual, selfish pursuits, rather than being paid or volunteering.

He further goes on to ask how to create a next generation, collaboratively-enhanced version of MapQuest? What features would lead people to naturally annotate maps, thus providing rich commentary and added value that you see in Amazon and eBay?

In response, it is argued that geographic data itself is more complex than hyperlinks links or online purchases. For MapQuest to become a platform, hooks are necessary to tie in your own data. These hooks are the very coordinates that (corporate or governmental) data providers like to keep as close to their vest as possible. Many of the internet killer apps had their handle problem worked out in advance ISBNs for Amazon, PNRs for Expedia, URLs for Google, and all of these identifiers are public. As geographic coordinates are difficult to remember for humans, and addresses difficult to handle by computers, mechanisms for auto-generating geo-tagged content, so-called geocoders, need more development.

Furthermore, geographic databases are shared among many organisations. They can't be crawled (Google) or centralised (Amazon) easily because no one organisation manages all geographic data, nor can they, for example for copyright reasons. Distributed content serving requires heavy-lifting aggregators and geocomputation is very expensive: a great opportunity for online mapping companies to add special value?

Also, geographic standards are required to leverage inter-organisational geographic data. The OGC specifications are too heavy-weight for many situations, where RDF is just fine. OGC does a good job of enabling complex things, but the technology doesn't scale down to make simple things easy. Have geoscientists been too proud to accept “good enough” technologies and are instead waiting for perfection? For example, the British OS always justify their pricing structures because they map with high accuracy all parts of Britain, also those areas that are not economically viable. However, does one always need the precision offered by OS? Furthermore, OGC specifications typically focus on the location aspect, ignoring other attributes of geographic features.

For collaborative cartography, people need to be able to trust the data they see. That means contributors have to be able to develop reputations and have to be able to see other peoples' reputations, e.g. through querying FOAF networks. These can be used as filtering tools to weed out “geographic spam”. Also, the data has to be served and advertised and queried easily. Better and ubiquitious positioning technology has to become available as well. But at the same time, there have to be measures to be put in place to guarantee one's privacy: people have to be able to hide their identity and whereabouts at will.

Going back to the idea of enhancing the current MapQuest service, the only way they presently provide a personalisation mechanism, is by setting a cookie to store user preferences during a user session. Other online mapping providers take this further, enabling users to sign up for personal services, such as MyMultimap and MyMaporama. Users can retrieve their personal settings during each following visit upon sign-in. Search patterns of signed-in users be analysed not only to improve online mapping services even more, but also to discover popular locations. They can store their popular restaurants and bars together with short reviews to share with their friends: “People who live near here recommend these restaurants...”

The stored information not only has to apply to locations, but can apply to routes as well: “Users who traveled between these waypoints have suggested these alternatives...”, for example corrections where the MapQuest route just didn't work or optimisations where there is a faster route, and some long series of scenic or theme based detours. It's just a small step, but it's certainly one step it the right direction!  permanent link for this entry

Nomen est omen?

Yet another phony press release or should we take this one seriously? The French online mapping provider Maporama sends out pres releases regularly, informing about their latest coverage enlargements and updates. This time however, the press release announced that Maporama takes a new start and becomes Maporama International.

For a while, Maporama was not considered a major player in the online mapping arena because of its weak financial situation. For one, it has been well known that Maporama paid Yahoo! Europe to be its mapping provider of choice. With the internet economy dwindling, Maporama was having sincere difficulties paying up. Furthermore, there have been licensing issues between Maporama and its incubator GeoConcept SA. Although the software licenses were well past their sell-by date, the GeoConcept software was still being powering Maporama's core technologies. A French court firmly ruled Maporama pays a daily fine for illegally using GeoConcept software. Not sure what the current situation is on either of the aforementioned.

Finding a new investor and chairman in Sydney Drahy seems to be a healthy development. The arrival of Sydney Drahy in the shareholding will bring us strong financial support and new perspectives of development. With turnovers in France, Europe, and the United States each making up one third of the total turnover, Maporama truly lives up to its new name “Maporama International”: nomen est omen?  permanent link for this entry

Wednesday, February 04, 2004

Mobilotopia

Having spent the best of last week in Vienna, Transmediale pulled me to Berlin. Special highlight of the art festival was the Mobilotopia conference panel featuring Marc Tutors, Ben Russell, and Drew Hemment. New to me were Jason Harlan (BlogMapper, RDFmap) and Teri Rueb.

The follow-up discussion and workshop to the MobiloTopia conference panel later on Monday night provided a platform for many more other artists to show their work. Some interesting ones are:

After these short presentations, the audience and panelists split into several groups each discussing one of the 4 topics:

  1. art
  2. social
  3. technology
  4. network

Jason and I were on the technology panel. Not too much interest: it's a lot to hope for at an art festival. The other groups were very popular indeed. Well, I am sure the locative media lab will find much more of a warm technology reception at ETcon in San Diego!  permanent link for this entry