MMC review and possible improvements

[HervePruvost]

Dear MMC-developers, this an issue started to collect some remarks and opinions from MMC users. Hope it can bring some support.

• in Linkmodel.xsd there are „k=“ for Key and „v=“ for Value. In Multimodel.xsd there are instead „key=“ and „value=“. Maybe be it could be harmonised putting "key" and "value" everywhere.

• XML-element "MultiModel" disposes of an id attribute ("uuid"). However element "LinkModel" has not.

[spacehorst]

Hi Herve,
sorry for the late comment but I just detected your question by accident.
Thanks for opening this topic. This is the right place for the future development of this schema.

1. names of the attributes
   We had long discussions abot this naming convention and ended up with the current state. The reason for the single-letter naming in Linkmodel.xsd is to keep the amount of data as small as possible. While there are potentially many more Link data sets then Model Elements in Multimodel.xml, we think this approach is fair. I hope this intention is stated clear enough in the XSD documentation.

2. Missing UUID attribute in LinkModel
   Well, this is a good point. Having an UUID there is not that expensive and would be more consistent. Structurally this is not explicitly required because each Linkmodel.xml is meant to be a containment reference from its Multimodel.xml. In other words, each Linkmodel can only exist with its parent Multimodel.xml.

I suggest to keep suggestion #2 up to the next service release of this schema.
Today we cannot say when this will happen. Seems the schema is doing quite well in practice :-)
As a new version would require a new voting process and a technical backward compatibility, we should consider to collect a representative amount of improvements.

[HervePruvost]

Hi Sebastian,
no problem. Thank you for your reply.
These were some remarks that have raised within a research project, as we were implementing multimodels with a partner for building energy simulation purpose.
These both aspects did not hinder the simulation process at all which worked quite well. The questions were more about the consistency in the vocabular. By the second point, we had the need to make references to LinkModels, so we made them by relative file paths pointing to LinkModels within the MMC.

If other remarks come in my mind, I would share them here.