[wdmmg-dev] Terminology

Wed Jan 19 14:04:56 UTC 2011

Hi,

We need consolidate our terminology, and I would recommend to go with already established OLAP standard. Maybe you all know internally what each part of the code does, but it is difficult for external person to understand it. Also for the future, if we stick with OLAP terminology that is being used "in the wild", we will not have to explain our data model and there are plenty of books out there for anyone to read, including wikipedia. Also if anyone ever comes to wdmmg source code, he would be able to understand it better, just by reading OLAP stuff on wikipedia.

In short:

There are two very simple concepts: measurement and context.

Fact - you measure them by measures, such as "amount"
Dimension - describe context, you refer them in text "by ..." or "for ..." such as "amount by region" or "average for january"

Drill-down - add dimensional detail, go from summary into details (no "break-down"), can be compared to expanding a directory in a file browser and getting all subdirectories

Currently used terminology and clarifications:

* Axes should be called dimensions
* "Entry" collection is a "fact collection" containing "facts" (no need to rename, just need to know)
* any other collection referenced from "entry" collection is a "dimension collection"
* any key (attribute) that is going to be used as context in reporting is a "dimension", that is every attribute that we are going to "report by" is a dimension
* any attribute that is going to be measured, reported, aggregated is "measure",currently we have only "amount"
* avoid using break-down instead of drill-down

There should be no difference from analytical point of view between: classifier, entity, place (future), ... all those are just dimensions. Only difference is, that Entity is "flat" with one level, and Classifier has more levels.

I've created a ticket for this:

	http://knowledgeforge.net/okfn/tasks/ticket/573

If we follow existing OLAP terminology, then we would be able to say that "we have built first international governmental budget datamart"[1] ;-)

Regards,

Stefan Urbanek

[1] http://en.wikipedia.org/wiki/Data_mart