There are two levels to this issue - methodological and pragmatic:
With experience, we have shifted during the years towards more explicit modeling rather than using static model properties. We have found that when we add data items and use their values to influence behavior, we get more flexibility at runtime.
For example, you can add a Boolean data item 'Mandatory' to any model and initiate it to either 'Yes' or 'No', and then you can change its value later depending on the circumstances (e.g. if it's mandatory under certain conditions but not under other conditions).
We are not planning at the moment to add a runtime "read property" capability, as it doesn't seem a high priority issue. Is there a specific case where you believe this is a must (and cannot be implemented as suggested above)?
For best results, use the Firefox browser..