When we originally designed and wrote Tessitura, there were some pieces of functionality that made their way into the system that I never really thought were fully formed or thought out. As a result these features tend to get very little use or confuse many users.
One such feature is constituency-based security. This feature allows you to define rights to one or more constituent functions (i.e. editing addresses) based on the highest ranking constituency of that constituent.
Development stories concerning constituencies are starting to come into view and we need to explore whether this security paradigm is worth continuing.
In my opinion, the whole idea of constituency-based security loses a bit of its shine as soon as you realize that it's all based on the highest ranking constituency only. In addition, the concept was put in place well before Control Groups was ever thought of and we might have thought twice about it if Control Groups were in the original plan. It's not that Control Group functionality replaces constituency-based security, but rather that sometimes you have to step back and wonder whether there aren't too many layers of complexity. Especially if one of those layers is not often used--a survey of Network implementation consultants seemed to show that very few sites ever implement this functionality, at least initially.
Another factor prompting this thread is that several times our internal discussion on on how we might better use constituencies going forward has been hampered when someone says "yeah, but don't forget about constituencies and security".
We've tried to model out what would happen if we changed the model so that security could be based on "any" constituency instead of "highest ranking" constituency. But that always creates problems. What if a constituent has both constituency 1 and constituency 2. Constituency 1 says they're allowed to edit addresses and constituency 2 says that they are not. What to do then?
So some help here, please. How many sites use this functionality now and how passionate about it are you?
As always, thanks in advance for the feedback!
We do use Constituencies for security, but it is very limiting. As already pointed out in many of the replies, you have the issue of only the highest ranking constituency being used to determine security, and of course, you have to make sure that the appropriate constituency is applied to the customer in the first place.
What we would find far more useful is something more logic-based, where we could define rules that automatically apply to customers, and as customers data changes, the correct rules automatically apply. We really need more elements exposed than just constituency for determining permissions.
For instance, the ability to create a rule such as:
Customer is a Subscriber for more than 2 years and a Donor over Level XXX, and has a value of YYY for Attribute ZZZ, then
User Group 1 can edit,
User Group 2 can view-only,
User Group 3 can see name in a list but no other details.