by akivela » Wed Mar 11, 2015 3:36 pm
Hello
Nice work. You have literally modeled the graph nodes and edges. Nodes are connected to edge topics that carry an occurrence text for distance and time. The idea is that distance and time are properties of the edge.
I have attached two alternative examples that are little simpler, in my mind.
First example (test_v2_1.wpr) models Connected locations with quad-associations. Association roles are Location, Connected location, Distance and Time.
Second example (test_v2_2.wpr) is similar to the first example but models Distances and Times separately. There are two different associations types: Distance and Time. Both association types have tri-associations. Addition to Location and Connected location roles the association contains either Distance or Time role.
What is weak in these alternatives, is the requirement for symmetric associations. If Amber is connected to Ankh-Morpork, there must be a separate association for reverse connection too. Yet another weakness (or strength, it depends) is the distance and time topics. It requires some additional attention when the author edits scalar value topic that may be player in several associations. What is definitely a strength in these examples, is the ease of editing/importing/exporting. One can easily import/export association tables from/to spreadsheet applications such as MS Excel and do some of the editing there. It took me only few minutes to copy stuff from your topic map to Open Office Calc and paste ready associations back to Wandora.
Deciding which way to go, requires a larger vision and is little a matter of taste.
Kind Regards,
Aki / Wandora Team
- Attachments
-
test_v2_1.wpr
- (2.91 KiB) Downloaded 14265 times