how compression works -- details

Any details on this? I was at TechEd and heard about it -- a bit. I understand the prefix portion of the compression algorithm and how its placed on the page. But what is the dictionary for? If I found a pre-fix that's common to 3 DIFFERENT values, would it change the prefix or migrate the value to a dictionary? Is the dictionary for cross column compression? When do prefixes and dictionaries get updated -- as soon as the data hits the page? Cool feature, just trying to see how it compares to Oracle.

I assume this will be free? I saw in Oracle 11g they have announced a new Advanced Compression feature, but they took it off the Web site for now -- that would make it $$ now, but it used to be free. Interesting....

Very good info here: http://blogs.msdn.com/chadboyd/archive/2007/07/28/katmai-sql-2008-data-compression-including-backup-compression.aspx

|||Very good indeed! Anyone know when we might see it in the CTP?

|||All I can say for now is 'a future CTP' (which ofcourse does not help very much).

We are still writing code for the compression feature at the moment, and are getting very close to being code complete.

When it is in a CTP, I will be more detailed in my responses.

Thanks,|||Any update on when we'll see compression in a CTP?

|||It won't be in the next CTP, but very likely in the CTP after that one.

Thanks,