[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: [ubl-lcsc] 9/29/2002: Normalization Concept
Tim, My comments on your Containership paper. Thanks. Monica -----Original Message----- From: Tim McGrath [mailto:tmcgrath@portcomm.com.au] Sent: Wednesday, September 25, 2002 9:05 PM To: ubl-lcsc@lists.oasis-open.org Subject: [ubl-lcsc] At the upcoming face to face we have a scheduled agenda item At the upcoming face to face we have a scheduled agenda item for discussing the ideas of containership based on normalization. To demonstrate what this means to our modeling, i have taken the current 'Party' structure and applied to principles given in the Containership position paper. Please find attached a PDF file describing the structures, before and after normalization. [NB in the process of this I made some analysis and design decisions - this was not just a mechanical process. the exercise did, i feel, expose some wekanesses and help design some better solutions] To show this is not just theory, i created an hierarchical tree view based on the context of Party as applied to an Order. The becomes the schema structure. To show this i am also including an XSD for this structure - but it is not the UBL implementation of XSD, just a basic and crude version to show the principle involved.(it also comes with a diagram courtesy of XML Spy). To prepare for our discussion i encourage you all to consider this material carefully. If anyone wants to try their own exercise feel free to do so. There are several implications for our future work in this. One of which is the use of the current spreadsheet for our logical model - how can we show the structures such as in the PDF file using this form? Is there a better tool? Another consideration would be, how does this concept fit with the idea of 'context' (e.g. look at the way 'contact' is now a repeated set)? Is 'context' somehow related to the links/relationships between containers? With respect to the CCTS naming rules, can we say that these ABIEs are the 'Object Classes' and their BIE components are 'properties' of their 'Object Class' ? If so, can the links/relationships be the property qualifiers? For example, the ABIE 'Party' has (potentially) multiple properties of 'Contact'. What differentiates them is the relationship (or context), such as Receiving Contact, Shipping Contact. That makes the Object class = Order, the property = Contact and the property qualifier = Shipping, giving us Order.Shipping Contact etc... I hope this is a useful primer and i look forward to the forthcoming discussions. -- regards tim mcgrath fremantle western australia 6160 phone: +618 93352228 fax: +618 93352142
Attachment:
position-mcgrath-containership-01_mm1_092902.doc
Description: position-mcgrath-containership-01_mm1_092902.doc
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Powered by eList eXpress LLC