I apologize for the potentially misleading title – many will likely assume this is about selling design but actually what i want to write about here it is more about buying design.

The idea for elaborating on this subject came from seeing the recently introduced Wikimedia maps, in particular its appearance at the lower zoom levels

which is a great example of map design going wrong and seems also well suited to explain the causes why this happens in maps today.

I should probably add upfront: i am not involved in the development of Wikimedia maps nor do i have any particular investment or interest in this project and although i point this out as an example of quality problems in map design here it is just an example and does not represent a case of particularly bad quality – you will be able to find many similar examples in other maps.

What the example above shows is quite simply bad map design. This is not limited to this particular map scale, you can still observe the same as you zoom in several levels but it is most obvious at this scale. One of the most basic rules of good map design is to give the map user a correct and consistent impression of reality. Note i do not say realistic here. A map is not a photograph but an abstract depiction – this depiction is however still meant to show reality. This means if you depart from creating a faithful depiction you should always do so deliberately and consciously and your aim has to be to avoid creating a misleading impression on side of the map user.

If a map design no more tries to correctly and consistently depict reality it essentially looses its raison d’etre and becomes a piece of pure art – if it qualifies as such – or simply a pile of digital junk.

Now at the scale of the map above this is not easy – there is not much room to show the complexity of reality but it should be obvious that you can do much better than showing the baltic sea as several unconnected lakes or Japan as a peninsula. The standard OSM map style – though certainly not great in that regard – is obviously doing a much better job.

As many of the readers probably know i am fairly familiar with this particular design problem, i.e. consistently representing coastlines and waterbodies at coarse map scales but i am not writing this to promote my work on this subject. Any approach to this problem is full of subjective choices and any map designer is free to use a different approach here but it is essential to actually make conscious design decisions and not just render the random results coming out of the arbitrary combination of various algorithms.

So much about the problem from the perspective of map design but my actual subject today is what causes these problems to prominently appear in maps intended for a broad audience.

Although Wikimedia is not a company it works quite similar to a commercial enterprise when it comes to executing a map development project. What happens is that the decision is made that a map is to be developed and people within the organization, hopefully mostly those competent on the matter, are tasked with coming up with specifications of the requirements for this map.

Here the problem usually starts because it is inherently difficult to put design aspects into a project specification. There are essentially two ways this is done:

• specification by example – the map design should be in certain aspects like some existing reference design. This method has two major problems: (a) it only works for standard products and not for new, innovative or uncommon designs and (b) in many cases – either due to internal policies or by law in cases of public institutions and involvement of tax money – it is not allowed to do this because the specification based on an existing product is not neutral.
• specification through technologies used – instead of specifying the design itself it is only specified what technological means are supposed to be used in creating this design. Essentially this is a bit like comissioning a portrait painting and specifying it is supposed to be painted using water colors.

Given the constraints of the first method the second approach is by far the most common method to address design aspects in map specifications. It is probably clear that practically this rarely ensures a certain level of quality in design and it can even be counterproductive in a lot of cases. In the Wikimedia maps example the technical requirement to use a vector tiles approach (which was likely part of the specification) for example leads to the need to massively reduce the data volume for the lower zoom levels which is done in a way that adversely affects the rendering results. Without knowing the actual specifications for the Wikimedia maps it is a safe bet that these do not contain any requirements regarding the consistency in waterbody appearance at all.

On a more general level the problem here is that maps are created to fulfill certain economic needs. Maps address these needs through design and design is enabled by technology. This mechanism is not specific to maps, it is universal to all fields of design. In case of modern digital map design however technology has a much higher significance on what is possible and can be efficiently accomplished than in other, more conservative design disciplines.

The problem is that this mechanism is often not respected in the way decisions are made – it is ignored that design is the central link allowing technology in maps to address economic needs and instead it is often assumed that implementation of technology somehow magically leads to fulfillment of the economic goals. Traditionally people say what is possible design-wise is usually limited by what can be done technologically but in map rendering i have the strong impression at the moment that often use of many readily available technologies is constrained by the limited founding available to design in map projects since budgets often essentially end at the technical level.

For solving this problem it probably helps to look how other fields in design deal with this matter. If you look at how projects are managed in the fields of architecture and industrial design you can see that drafts have a much higher importance there. They can range from simple concept sketches to detailed design plans. It is common to make design decisions based on competing drafts by several designers developed independently to ensure a wide range of possibilities is taken into account and to give new and innovative ideas a chance.

I have never seen this happening in a commercial digital map design project. It is of course common to try different design variations during development but having several completely independent drafts made would be extremely unusual. You can of course argue that architecture and industrial design with total budgets often several orders of magnitude larger can afford paying for several drafts that are not used in the end while map design projects cannot – there is certainly some truth to that but even in those fields people usually do not have money to burn but spend the money in ways they think will ultimately turn out to be good investments.