Modeling LID

From lid.netmesh.org

Jump to: navigation, search

Contents

{proposed} This is the rough beginning of a project to build a model of identity data. This project is the first step in a project to model LID.

When this project reaches a stable state, this page will provide an overview of that model and links to detailed specification of that model. For now, these pages amount to thinking out loud. So, some of these pages violate the guideline that wiki pages should be short. Wiki enables your feedback and ideas to be entered right here, by changing or replacing my text. Or by entering text at the discussion tab.

Please contribute


Modeling LID--Planning

The first goal is a generic model of identity data. The first step toward that goal is a model of vCard data. A wordy discussion of the thinking that led to the current draft: Modeling LID--Planning

According to plan, the first step was to model vCard data, test the model, and correct. The next step is to model FOAF data.

Modeling LID--Test first

For starters, i will test the model by attempting to use it to model each of the example vCards in RFC 2426.

The next tests will use the example FOAF documents in FOAF Vocabulary Specification.

Then i (we, i hope; please contribute) will plan some real testing. The current draft: Modeling LID--Test first

The current draft model is on three pages:

Modeling LID--The classes and associations of the model

The initial model contained a small number of classes and associations, built by looking at the vCard spec, with an eye on the goal of a generic model of identity data, however specified.

The current draft: Modeling LID--The classes and associations of the model

Modeling LID--The objects of the model

Because the model we are heading for is designed to be changed and expanded by adding {objects | data [take your pick]}, much of the model is presented as objects (and links), rather than classes. If the model is successful,the classes and associations will be generic and everything specific to vCards will be included in the model as objects and links.

In preparing this model, some lurkers were discovered. The first of these is place, pointed out by Johannes, always on the lookout for meaning, rather than just data. The second, lurking in vCard, but explicit in FOAF is person.

Place is tricky at best: Consider post office caller numbers and fleet post office addresses. Person is straightforward, but raises a hard modeling question: Since we are developing a model of LID, should Person be a class, or just another object?

The current draft: Modeling LID--The objects of the model

Modeling LID--Testing the model

Testing the model means using the model with the intent of demonstrating an error in the model. Before doing real testing, i'll first take some vCard examples from RFC 2426 and represent them using the model. Then some FOAF examples, from FOAF Vocabulary Specification. This is, after all, what passes for testing in some shops. And it will have the excellent benefit of explicating the model by providing examples. The current draft: Modeling LID--Testing the model

Modeling LID--Repeat until satisfied

After executing the test plan, let's regroup and iterate. The current draft: Modeling LID--Repeat until satisfied

Please contribute. Thanks.

Joaquin Miller

Personal tools