I'm not sure whether it is possible to represent a VMS with several possible display layouts. Some full graphic displays have a number of predefined layouts with specified text and pictogram areas, e.g. a) 1 pictogram on the left, 3 lines of text on the right; b) 3 pictograms on top, 1 line of text on the bottom. I guess you could define all possible elements and just use the ones needed for a given message?
I also didn't find position info for text elements - is it even possible to have more than one text display area on a VMS?
I only found the possibility to show several pages of text.
Multiple text elements with specified positions would be absolutely necessary in my opinion (example again: http://images.google.de/images?&q=dwista).
I would add to the discussion that the modeling team had in mind to cover needs starting mostly for VMS managed in Mare Nostrum that are VMS along the motorway compound of one text area and one or more pictogram. (for detaails see the presentation made in Berlin 2010 user Forum http://www.datex2.eu/user-forum/1_Paoletti_VMS.pdf )
In the current model different text areas are seen as as different VMS. the central unit could be the same or distinct units could manage all VMS.
of course this has been driven mostly by our experiences of real VMS managed, it should be useful to best model the cases to know more details of how the VMS is implemented in order to see if the concept of VMS is correctly shared among us.
In the case where a VMS has dynamically configurable display areas as you describe then you can set the "dynamicallyConfigurableDisplayAreas" attribute in the VmsRecord of the VmsTablePublication. You could also set the default (or normal) configuration via the VmsRecord.
If the default configuration is active then the VmsPublications does not need to say anything more about the configuration. But if the configuration is not the default the actual current configuration would then have to be given in each VmsPublication which would override that given in the VmsTablePublication. I realise this is not very efficient, but the alternative would be to introduce some indexed set of configurations in the VmsTablePublication which could be referenced in the VmsPublication - this is possible but it would introduce yet more complexity into an already complex pair of publications. This probably needs further discussion.
Until now the premise for this pair of publications is that each VMS has only a single area for the display of text, whether that is for displaying a single page of text or a number of sequencing pages of text. If there is more than one area for displaying text then it would have to be modelled as separate VMSs. If that is a severe limitation then we will need to do some remodelling.
I would start with adding position attributes and a textDisplayAreaIndex to vmsTextDisplayCharacteristics and to allow more than one text area (so text areas are treated similar to pictogram areas). This would be needed for more complex VMS anyway in my opinion (see also http://www.datex2.eu/content/vms-static-info-dimensions-background-image...).
With that it could be possible to define all text and pictogram display areas for all possible configurations and later on just use the ones needed for the actual configuration. It would not be absolutely necessary to explicitly define the allowed "configuration sets" because we do not want to control the VMS.
Thilo,
I have now created an updated VMS publications model which addresses all your concerns.
It allows for multiple text display areas on one sign, a "painted" static background (by reference to a URL), the absolute coordinate references (in metres) of the positions of each display area on the overall sign and the height and width measurements of each display area.
This will allow the sign information to be exactly replicated or simulated by the recipient for display in their system, confident in the knowledge that it will be an exact representation of what drivers are seeing. This seems to be the essence of your requirements.
I shall promote this updated model in the TC and TC278 WG8.