January 29, 2009
This is a really big distinction but is only mentioned in passing in BOL (check here), but approximate numerics are just that: approximate. Someone who understands the storage engine better can explain why a foat or real can’t store an exact number, but they can’t and don’t. You might put 1.5 into it and later come to find that the value is 1.4999.
If you need an exact representation of a number that isn’t an integer, you need to use numeric or decimal. These are functionally the same, so it doesn’t matter which you choose.
January 23, 2009
Chris Webb has the scoop here: http://cwebbbi.spaces.live.com/Blog/cns!7B84B0F2C239489A!4133.entry
What a shift in the Microsoft BI stack. You can’t show the BI Burger to your clients anymore. Maybe there was not enough demad for the planning peice for Microsoft to justify continuing the development and sales effort around the product. Now we have yet another example of an acquisition that soon was re-worked and then kill by the acquirer.
I have also heard rumors that Gemini might be something to take the place of the planning piece. You’d need some sort of workflow engine, but maybe it could work. We’ll see.
Also, the layoffs affected the sales area in BI significantly so MS is definitely cutting their loss in BI. I agree with Chris that it is important for Microsoft to publish some sort of new roadmap. It makes it hard for partners to invest in new Microsoft products if they kill them off so soon after launch.