The external hardware is decribed by means of a configuration file in JSON format. Every available hardware must be grouped into device, subdevice and channel. A device could be a data acquisition card, an external FPGA, or a pointing device such as a mouse. The subdevice denotes a certain function such as a group of digital-to-analog outputs or a bank of digital inputs. Finally, a channel is one of such a group of inputs or outputs. The complete list of available keywords can be found at Available Keywords.
Let's consider a first example

{
  "device0": {
    "library": "libcomedieeros.so",
      "devHandle": "/dev/comedi0",
      "subdevice1": {
        "type": "DigIn",
        "channel0": {
          "signalId": "emergency"
          "inverted": true
        }
      },
      "subdevice5": {
        "type": "AnalogOut",
        "channel0": {
          "signalId": "aOut0",
          "scale": [ { "id" :    "aOut0",
                       "minIn":   2000, "maxIn":  64000,
                       "minOut":-9.787, "maxOut": 9.923 },
                   ],
          "range": [ { "id" :   "aOut0",
                       "minIn":      0, "maxIn":  65536,
                       "minOut": -10.0,	"maxOut":  10.0 },
                   ],
          "safe": 0.0,
          "unit": "V"
        }
      }
    }
  }
}

Every element is composed of a key / value pair. The key must be unique for every hierachy level. The key for devices, subdevices, and channels contains the number of the device, subdevice, or channel, respectively.

A device must contain a library key. The system can then search for the necessary library. This library must be present in the system and will be loaded dynamically upon parsing the configuration file. For available libraries check Hardware Libraries

A subdevice must contain a type key. This type must coincide with the subdevice type of the hardware library in use. Some subdevices contain channels of different types, e.g. a subdevice might hold several digital i/o's. Some will work as inputs while some other will be outputs. In this case you define the subdevice as being of type DigIn. A single input channel inherits its type from the subdevice. If it is an output you can redefine the type for the channel.

  "subdevice5": {
    "type": "DigIn",
    "channel0": {
      "type": "DigOut",
      "signalId": "enableDrv"
    },
    "channel1": {
      "signalId": "readySig1"
    }
  }

Every channel must contain a signalId key. This is an indentification string. The HAL manager registers the channel under this name and you have to request an input or output from your EEROS application with this name.

Channels carrying an analog signal or channels with counter inputs must be assigned a scale and range key. With these values the HAL is able to translate between the analog signal and its digital value as this digital value is necessary when accessing the hardware converter (counter, adc, dac). For examples in how to setup the scale and range properly see Scale and Range of Analog Signals. There are cases where the HAL should account for several stages on a scalable input or output, see Setting Multiple Scales and Ranges.

An analog output can have an associated safe value. If a NAN of INF value would be written to its output, the output value will instead be set to its safe value. This helps to guarantee that a drive will not get out of control. NAN or INF can occur, if a connection between blocks is open, if a block does not run or if an algorithm within a block produces this result due to an error. If the configuration does not specify a safe value, it is set to NAN.