When publishing data sets from different sources, it is important to agree on a common naming scheme. Otherwise, comparison is impossible. At the moment, a number of existing proposals to name CN components exist:
(Inserted by Bart) This is the current proposal of Funkfeuer to name entities in datasets, as documented on Github:
By node we denote a physical location where one or multiple devices are grouped together and are directly interconnected on layer 2. A node is not to be confused with a device.
Strictly speaking, the nodeDB is the set of different registries (IP address, nodes and device registry and person registry). There are other components (as mentioned in the introduction) which interplay with the nodeDB. For example the map, the dashboard, etc. These components usually query the nodeDB via a REST interface.
By device we mean an individual piece of (router) hardware which is an independent participant in the mesh network.
By dashboard we denote the main user interface through which a user of the community wireless network can see "his" community wireless network. It also gives the user a comparison actual data of the network (statistics, reachability etc) with the stored data in the nodeDB
Guifi also has a method to name components in their CNML data, a short overview is given on their wiki:
<devices> cpu, memory, storage, devices, radio level etc. </devices>
<mapping> gps, name </mapping>
<owner of device>
<scale of interest to the network>
<personal keyring>(hold private, only given volountary)
A more correct format is described in the CNML overview.