Thanks Lars and Jon for your valuable input. Now I would like to summarize our discussion herein:
As of HANA SP11,
1) Querying a HANA view with star join with any join will always result in execution of fact table part. The effect is even if one query an attribute of a dimension, it will result in execution of fact table and not dimension table only.
2) For the consumption in BI, its desirable the querying a dimension should result in execution of only dimension view and not fact. The effect is LOVs runs slow un-neccesary. Refer how a star schema operates say in universe.
3) All use cases are possible, see all dim values or only those related to facts only. However, star join, as it seems to a BI developer, will seems to be a regular star schema where querying a dimension should result in only querying dim and not fact. In BEx as well, its the same.
4) In absence of above, I don't see how well HANA view can be directly consumed in BI tool.
5) In my opinion, SAP HANA/BI front end dev team has to think better in improving the semantic layer and integration with BI tools where integration is not name sake but support existing workflow in various data sources. Otherwise it leads to various patchwork, workarounds resulting inefficient and less maintainable BI.
6) With regards to point #6, it seems to me a query tool of kind of BEx query is need which is more OLAP aware (than Universe) and in addition may be universe can also be better integrated.
Thanks once again.
Regards,
Nitin