we're talking about structured lists. You might wonder: How do I make
these structured lists? In this post I will try to explain how to do it.
It's actually not that difficult to create :-)
I will take Switzerland as an example. Their system is described at
We're going to create or convert tables. For this we need to know the
fields to include. Some important fields:
* id - the unique id assigned by the local registry, this is the primary key
* name - name of the object (or a description)
* address - the address of the object
* municipality - the municipality in which the object lies
* lat - the latitude
* lon - the longitude
* image - image of the object
You can of course add more fields and the name of the fields should
probably be in the local language.
The fields I used for the Swiss monuments are:
* CH1903_X (they use a strange lat/lon system in Switzerland)
* KGS_nr - this is the unique id
Now we need to make two templates
* A header template which is going to be at the start of each table
* A row template. One row per monument containing all the information.
These templates make the lists look pretty and make it possible for a
bot to harvest the information (more about that at
For Switzerland I created
Now we got the base of the template system. If you don't have any lists
yet you have to get a dataset to start from scratch. If you already have
lists you need to convert them. If you're lucky some bot operator is
able to convert a lot of the lists automagicly with complicated regular
expressions. With the Swiss lists I was able to convert quite a lot with
a bot, see for example
(check the source and history).
The remaining items need to be fixed manually, you should mobilize some
users to help out.
Now you have structured lists! This is where I stand now with the Swiss
project. I will do a follow up most when I got some of the nice tools
working so we can actually make use of these structured lists.