To whom it may concern,
while in Regensburg looking into possible cooperation with my research
late last year, I noticed there's a slight issue with the ER diagram of
the database model in the MWB file which concerns the cardinalities of
the relations between tables.
For example, the relation between *issue *and *person *(relationship
*issue_asignedTo*): on the person side, there's *0..1* cardinality,
which is correct because it says "an issue can have 1 or none
assignees". However, on the issue side there's *1..** cardinality,
suggesting "every person has to have at least one issue assigned",
whereas the right cardinality should be *0..** meaning "a person can
have multiple or no issues assigned". At least that is my understanding.
If so, the inaccuracy is spread througout the model, since there are
very few*0..1* and *0..** cardinalities and, I assume, in many instances
there should be.
Nevertheless, the issue is only with the ER diagram of the model and
doesn't affect the database structure itself (since the setting of the
*Mandatory *checkbox in *Foreign Key* tab in the *Relationship *settings
in the MySQL Workbench - which affects precisely this - is not reflected
in the SQL export) nor the functionality. The same goes for changes
potentially made on the cardinalities (going from *1* to *0..1* and from
*1..** to *0..**). The change would therefore be cosmetic only.
But Wolfgang's opinion was that the ER diagram should represent the data
model as accurately as possible (even more so when the changes don't
affect anything else) and that I should write an email about it to this
mailing list for collective consideration.
With regards,
Petr Pícha
