How the Model Registry Works
This is a bit of an experiment. The idea in a nutshell is to provide a means for folks to be able to register architecture models that are placed as open source in the public domain so that others can find them and use or add to them.
Why Do This?
There are many problems facing those that might want to collaborate in architectural modelling not the least of which are:
- exchange standards used by modelling tools
- intellectual property, security or commercial restrictions
- representation / consistency
- semantics of the elements used / consistency
- finding suitable models
On the face of it therefore nothing will ever get shared on this basis. We therefore decided to start with a different premise, namely that:
- folks want to share
- there are models that can be placed in the public domain under an open source license
which only leaves the problems with
- tool interchange standards (which we can’t do anything about other than identify what does/does not work and what suffers)
- semantics (which can be declared in views or any dictionary-like mechanism)
- consistency of representation (we can document, develop and declare design patterns but can’t solve major incompatibilities)
This leaves the problem of finding models - one of discovery which we can very definitely help with.
How It Might Work
The registry will provide
- a title
- the purpose for which the model was developed
- the framework used e.g. MODAF, DODAF, TRAK, NAF
- some means of indicating coverage / completenes - need to know the perspectives / viewpoints included
- a description
- the open source repository location
- the format / standard in which the model is held e.g. XMI version
- the url of the RSS feed for the model (so that interested parties or users can subscribe and be kept up to date)
The registry provides a means therefore to lodge a description of and for others to search and find models that might be suitable for use and extension.
Interested parties can keep up to date
- with new registry entries - by subscribing to the registry RSS feed
- with any individual model changes - by subscribing to the remote model RSS feed (most open repositories like Sourceforge and Google provide these as part of the package and for free).
It is assumed that some form of open source license would be used. The good thing about open source licenses is that full attribution is preserved and together with a change record and version control this is a pragmatic way of providing credit and model history.
It is almost certain that other documentation needs to accompany any model placed in the open:
- dependencies on other models / boundary
- scope / description
- some visual description
so that it is possible to make an assessment of whether it is worth using (as XMI isn’t very human-friendly). This is partly why this is being set up to empirically figure out what works and what doesn’t.
It’s not perfect but it’s certainly a lot better than the current state of play!
A dummy entry has been made in the repository. Improvements will be made by adding tags to each registry entry. It will also be possible to comment and rate each model, sort by all of these etc.
I’m working on a model of CONTEST - the UK Counter-Terrorism Strategy - because it includes common and re-usable parts, actors and because it is a common thread behind security and policing which often crops up. This will be placed into an open source repository and a registry entry made.
Of course this service will only be available to site members.