Timeline for UML Class Diagram: How can I represent "orthogonal" generalizations (or multi-inheritence)?
Current License: CC BY-SA 3.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 18, 2014 at 9:57 | comment | added | gwag | @Oliver: This is not the characteristic distinction between aggregation and composition. Rather, an association is a composition when it represents a part-whole relationship where the parts cannot be shared among two or more composites. See also this answer [stackoverflow.com/a/25120476/2795909] to a related question. | |
| Aug 5, 2014 at 8:46 | comment | added | Oliver Watkins | if it can exist without a backup then you are correct. It should be aggregation instead of composition :) | |
| Aug 5, 2014 at 8:13 | comment | added | Renaud Tarnec | I have an extra question please: shouldn't it be an Aggregation instead of a Composition? The Backup System can "exist" without the backed up one, isn't it? | |
| Aug 5, 2014 at 7:58 | vote | accept | Renaud Tarnec | ||
| Aug 5, 2014 at 7:57 | comment | added | Renaud Tarnec | @Oliver Thanks for your reply! You are right, with composition it should be fine. Actually, since all Systems are not back up ones I thought I would represent this specific case of a System through a sub-class... where actually it can be done with a 1 - 0..* composition | |
| Aug 4, 2014 at 18:16 | history | migrated | from stackoverflow.com (revisions) | ||
| Aug 4, 2014 at 16:18 | comment | added | umlcat | (+1) Agree, the "backup" feature "sounds" more like composition ... | |
| Aug 4, 2014 at 15:56 | history | answered | Oliver Watkins | CC BY-SA 3.0 |