I love seafood and was looking forward to some at a restaurant on the “Rocks” last week in Sydney, Australia that is. However, one of my TM Forum member pals and his wife convinced me to try a French restaurant with them. And, I am glad I did…a wonderful three course taste of France, starting with Coquilles Saint-Jacques, then Steak au Poivre, and a desert of flaming crepes (Crepes Flambe). The restaurant is “Chez Maurice and Linda”. Chefs Maurice Migneron welcomed us into the restaurant and gave us the choice of tables, and his Vietnamese-born wife is Linda served at our table. Take a look at http://maps.google.com/maps/place?rls=com.microsoft:en-us:IE-SearchBox&oe=&um=1&ie=UTF-8&q=Chez+Maurice+and+Linda+Sydney&fb=1&gl=us&hq=Chez+Maurice+and+Linda&hnear=0x6b129796569001c5:0x4017d681632a860,Sydney+NSW,+Australia&cid=10194646804675417769
Over dinner part of our conversation was about the Information Framework’s (SID) many uses. My pal stated that the SID should be focused on its use as the basis for interface development. When I replied that the SID can also be used as a starting point for data base development, the sparks started to fly. And, they could have been used to ignite the flaming crepes!
It is an interesting discussion, and one in which I am involved with some TM Forum Interface program members, including Marc Flauw of HP, our newest TM Forum Fellow. He and I are engaged in in the review of the Alarm entity that is being added to the SID. My first thought was that with 30 attributes, its apparent flattened structure was primarily suited to interface development and not data base development.
I won’t go into all the details of our back-and-forth discussion, but will provide a few examples along the way here. One thing I brought up was that it appeared a threshold entity was embedded as a set of attributes in Alarm. We did agree that this was most likely true, but from an interface perspective it would be treated as a data type embedded within the Alarm. This led to the identification of a transformation feature that would enable related entities to be embedded as a data type within an entity. Another thing I brought up was that it appeared there were quite a few associations to Individuals/Organizations playing various Roles (PartyRole). For some we agreed; for others the attribute was merely a reference that was not validated. This led us to agree that if validation (via referential integrity) was required for a data base implementation, the attribute would become a foreign key (or association) for a PartyRole.
The bottom line…there are transformations that need to be made when putting the SID to either use. These transformations often depend on its use and also on the rules that need to be enforced by the resultant model. And, hopefully there can be some automated support for some of these transformations, and that the configuration of the transformation can be stored for later use/updates. We are also working on this, which will show up as documentation associated with the import tool used to move SID entities into Tigerstripe (our interface specification and generation tool) and as documentation associated with how to transform the SID into a data model, which I am working on this year.
Off to China again next week, then back for a few days, and then off for a much-needed two week vacation. So, I’ll be taking a bit of a “blog holiday” until I return. Hope to see many of you at Team Action Week Baltimore in July!
Posted
06-16-2011 8:42 PM
by
John Reilly