Architecture Notes

We are opting for a classic XML publishing system architecture, layered as so:

• Presentation layer (HTML, PDF, RSS etc.)
• Processing layer (XML/XSLT)
• Storage layer (file system / dbms / archives)

This architecture allows various design problems to be isolated, and interfaces between the resulting components to be defined more cleanly. An HTML front end can be designed differently from the processing architecture behind it. And a storage/access model can be tuned to the requirements, instead of being imposed by a particular processing model.

On the lowest tier, we can begin with a basic CMS and grow into it.

On the middle tier, we can proceed with XML and its transformations.

On the upper tier, we can both present, and call out to, hypertext content in any format current browsers and readers can deal with.

The Framework Notes document has more to say about how these map to DHQ activities.

— WendellPiez - 10 Aug 2005