Proposal: remove of type attribute from Link object

In the wake of the NML sessions (for which the meeting minutes will follow shortly), Jeroen, Jason and I had some discussion on the schema. Based on these discussions, I propose two changes to the schema. Here is the first. I propose to remove the "type" attribute from the Link object. We currently define two values for this type: <link type="crossconnect"> and <link type="link">. There seem two reasons to distinguish between Links and Cross-connects: 1. Tie a cross connect to a particular switching matrix, so it is clear which switching matrix is providing the transport service. 2. Convey information on the volatility of a link object. A <link type="link"> is pretty static, while a <link type="crossconnect"> is more dynamic, and a client may want to query the status of cross connects more often. It is my opinion that there are better ways to convey the above information, rather then with the 'type' attribute. (Perhaps using a relation between Link and SwitchMatrix and an extension of the Lifetime object). Even thought these proposals have not been made, I think that removing "type" will avoid undeserved expectations. FReek

I second the motion; -jason On 9/21/11 12:15 PM, thus spake Freek Dijkstra:
In the wake of the NML sessions (for which the meeting minutes will follow shortly), Jeroen, Jason and I had some discussion on the schema.
Based on these discussions, I propose two changes to the schema. Here is the first.
I propose to remove the "type" attribute from the Link object.
We currently define two values for this type:<link type="crossconnect"> and<link type="link">.
There seem two reasons to distinguish between Links and Cross-connects: 1. Tie a cross connect to a particular switching matrix, so it is clear which switching matrix is providing the transport service. 2. Convey information on the volatility of a link object. A<link type="link"> is pretty static, while a<link type="crossconnect"> is more dynamic, and a client may want to query the status of cross connects more often.
It is my opinion that there are better ways to convey the above information, rather then with the 'type' attribute. (Perhaps using a relation between Link and SwitchMatrix and an extension of the Lifetime object). Even thought these proposals have not been made, I think that removing "type" will avoid undeserved expectations.
FReek
participants (2)
-
Freek Dijkstra
-
Jason Zurawski