Aleksey Studnev (1) [Avatar] Offline
#1
First of all thank you for the book - i read it with pleasure and it was extremely helpfull

However, i was very dissapointed with mondrian 4. It looked like a better alternative to mondrian 3, and definitely a big step forward in terms of architecture, especially what concerns schema, aggregate support, etc.

In practice all that does not look so good. Specifically, i met with a bunch of problems in mondrian 4, namely:

1) aggregates are not used when they should. It means that when aggregate is defined in the cube, it will only used to load segment, but not a tuple.

2) the code around aggregates, in mondiran 4 simply commented, and there is no development in this area.

3) the quality of the code for mondrian 4 looks very poor, comparing to mondiran 3. i mean that architecture should be changed, but it was stopped at some point, and the trails of old architecture still appear ( like virtual cubes, aggr stars, etc...)

4) vertica not supported in 4.X, need a patch to work

All these could be a problems of earlier development, but mondrian 4 exists for a long time, and new releases ( 4.0, 4.1, ... 4.3 ) generated with no visible enhancements in these areas.

I got impression that mondrian 4 development stalled and it is better to switch back to monrian v 3.X.
What do you think?