Are these the relationships that you are picturing?
People::__pkPeopleID = People_Join::_fkPeopleID1
People2::__pkPeopleID = People_Join_fkPeopleID2
where People and People2 are Tutorial: What are Table Occurrences? with the same data source table?
And if you use a portal to People_Join on a People layout to link "John" to "Fred" with "father" as the relationship, you want "john" listed as a relative of "Fred" when you pull up his record?
That can be done, but it requires two records in the Join table with the ID values swapped. A script can be used to automatically produce the "mirror record" in the Join table.
Thank you for reply.
I was wondering if it could be done in one record of People_Join, instead of two, using a field with two repetitions for the peopleID and two repetitions in the description of the relationschip-field in People_Join, with logic to depict the second relationship-description if People is related to the second repetition of PeopleID? (or any other solution using one record instead of two.
This way the chance of letting a relationship exist as the counterrelationschip is deleted, is reduced. This is sometimes possible if scripts are not good performed in web-solutions.
As long as you do not use script triggers this script should work equally well in web published solutions as it does not require any steps that are not web compatible.
Whether you use two records, two repetitions or two sets of link fields, you'd need some way to automate recoring the second pair of matching values. It's certainly an interesting idea and one that you can experiment with to see how it works.