The power of cognos cubes and AS cubes is completly different!
AS provide the MDX language which allow a user or developper to do
"anything" with the cubes.
you can create complex calculations which is not possible with cognos.

second point, cognos cubes can only be viewed with cognos products ! you
can't use any third party front end tools. For AS you can find a lot of
tools and choose 1 or more regarding your needs and usage (at least excel
can be used) (also, cognos can use AS cubes)

from an administrator side, I prefer AS rather then Cognos.
first because I have always the right numbers in a AS cube. with cognos its
a knightmare to have the same total (in the database and in the cube)

second: the crashes... in the past (I known cognos has work on it) if my PP
transformer application crashed before my save, then my entire cube become
corrupt!!!!

other point: the security and the performance of the cognos cubes is poor
and you can't know what you have to optimize in your cube. AS provide some
statistics and help you to add aggregations to improve the performance.

also, AS allow you to create new cubes based on existing dimensions (or
virtual cubes based on existing cubes) , this reduce the time to create new
cubes.
For example, today you create a employee contracts cubes, tomorrow you'll
create a nw cube for the activities, the dimension is already created and
ready to use. and if you want to cross-analysis your 2 cubes, simply create
a virtual cube and its all ! no space required on the disk.
With cognos, to do the same job you have to create 3 different cubes you
have to create 3 times the dimension and 3 cubes to process.

cognos has 2 advantages:
1. it create MDC file, so you can send it by email to your users
2. the process is done if your data source is dirty. (some times its an
advantages, but some other times its not)

For example:
imagine your contract fact table contain an employeeid "555" but this
employeeid doesn't exist in the Employee dimension, in AS your cube will
fail to process (or ignore the error regarding some process option choose)
In cognos, the process simply create a new employee in the employee
dimension, but, because there is no name associated, you 'll see only "555",
you don't loose any values, but your dimension is now dirty.

to finish AS support parent-child dimensions, custom rollup formula, cell
level security, , actions, write back, nultiple storage mode (MOLAP, ROLAP,
HOLAP, Real-time), dynamic formatting (changing the color of a cell
regarding a value; example: sales <500$ in red) and a lot of OLAP formula.

AS 2005 will provide new features like:
* multiple drill through
* new dimension management system (which provide reach end user experience)
* multi-currency
* full multilanguage
* automatic generation of the datamart from your cube design
* predefined cube templates
* visual source safe support

I hope this help you.

Alex,
please elobrate on the so called advantages thet PP has over AS

Reg

P.S. an interface does not make a better product ( if this is your selection
criteria i dont want to be the devloper makeing cubes in your company)